Merge remote-tracking branch 'origin/main' into refactor/permissions-manager

management/server/account/manager.go

@@ -1,6 +1,6 @@
 package account
 
-//go:generate go run github.com/golang/mock/mockgen -package account -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//go:generate go tool mockgen -package account -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
 
 import (
 	"context"
@@ -13,6 +13,7 @@ import (
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	nbcache "github.com/netbirdio/netbird/management/server/cache"
 	"github.com/netbirdio/netbird/management/server/idp"
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
@@ -22,6 +23,7 @@ import (
 	"github.com/netbirdio/netbird/management/server/users"
 	"github.com/netbirdio/netbird/route"
 	"github.com/netbirdio/netbird/shared/management/domain"
+	"github.com/netbirdio/netbird/shared/management/networkmap/nmdata"
 )
 
 type ExternalCacheManager nbcache.UserDataCache
@@ -61,13 +63,15 @@ type Manager interface {
 	GetUserFromUserAuth(ctx context.Context, userAuth auth.UserAuth) (*types.User, error)
 	ListUsers(ctx context.Context, accountID string) ([]*types.User, error)
 	GetPeers(ctx context.Context, accountID, userID, nameFilter, ipFilter string, all bool) ([]*nbpeer.Peer, error)
-	MarkPeerConnected(ctx context.Context, peerKey string, connected bool, realIP net.IP, accountID string, syncTime time.Time) error
+	MarkPeerConnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error
+	MarkPeerDisconnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64) error
 	DeletePeer(ctx context.Context, accountID, peerID, userID string) error
 	UpdatePeer(ctx context.Context, accountID, userID string, p *nbpeer.Peer) (*nbpeer.Peer, error)
 	UpdatePeerIP(ctx context.Context, accountID, userID, peerID string, newIP netip.Addr) error
+	UpdatePeerIPv6(ctx context.Context, accountID, userID, peerID string, newIPv6 netip.Addr) error
 	GetNetworkMap(ctx context.Context, peerID string) (*types.NetworkMap, error)
 	GetPeerNetwork(ctx context.Context, peerID string) (*types.Network, error)
-	AddPeer(ctx context.Context, accountID, setupKey, userID string, p *nbpeer.Peer, temporary bool) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, error)
+	AddPeer(ctx context.Context, accountID, setupKey, userID string, p *nbpeer.Peer, temporary bool) (*nbpeer.Peer, *types.Network, []*nmdata.PostureChecks, bool, error)
 	CreatePAT(ctx context.Context, accountID string, initiatorUserID string, targetUserID string, tokenName string, expiresIn int) (*types.PersonalAccessTokenGenerated, error)
 	DeletePAT(ctx context.Context, accountID string, initiatorUserID string, targetUserID string, tokenID string) error
 	GetPAT(ctx context.Context, accountID string, initiatorUserID string, targetUserID string, tokenID string) (*types.PersonalAccessToken, error)
@@ -106,8 +110,9 @@ type Manager interface {
 	GetPeer(ctx context.Context, accountID, peerID, userID string) (*nbpeer.Peer, error)
 	UpdateAccountSettings(ctx context.Context, accountID, userID string, newSettings *types.Settings) (*types.Settings, error)
 	UpdateAccountOnboarding(ctx context.Context, accountID, userID string, newOnboarding *types.AccountOnboarding) (*types.AccountOnboarding, error)
-	LoginPeer(ctx context.Context, login types.PeerLogin) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, error)                       // used by peer gRPC API
-	SyncPeer(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error) // used by peer gRPC API
+	LoginPeer(ctx context.Context, login types.PeerLogin) (*nbpeer.Peer, *types.Network, []*nmdata.PostureChecks, bool, error)                    // used by peer gRPC API
+	ExtendPeerSession(ctx context.Context, peerPubKey, userID string) (time.Time, error)                                                          // used by peer gRPC API for ExtendAuthSession
+	SyncPeer(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*nmdata.PostureChecks, int64, error) // used by peer gRPC API
 	GetExternalCacheManager() ExternalCacheManager
 	GetPostureChecks(ctx context.Context, accountID, postureChecksID, userID string) (*posture.Checks, error)
 	SavePostureChecks(ctx context.Context, accountID, userID string, postureChecks *posture.Checks, create bool) (*posture.Checks, error)
@@ -117,15 +122,16 @@ type Manager interface {
 	UpdateIntegratedValidator(ctx context.Context, accountID, userID, validator string, groups []string) error
 	GroupValidation(ctx context.Context, accountId string, groups []string) (bool, error)
 	GetValidatedPeers(ctx context.Context, accountID string) (map[string]struct{}, map[string]string, error)
-	SyncAndMarkPeer(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error)
+	SyncAndMarkPeer(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*nmdata.PostureChecks, int64, error)
 	OnPeerDisconnected(ctx context.Context, accountID string, peerPubKey string, streamStartTime time.Time) error
-	SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error
+	SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error
 	FindExistingPostureCheck(accountID string, checks *posture.ChecksDefinition) (*posture.Checks, error)
 	GetAccountIDForPeerKey(ctx context.Context, peerKey string) (string, error)
 	GetAccountSettings(ctx context.Context, accountID string, userID string) (*types.Settings, error)
 	DeleteSetupKey(ctx context.Context, accountID, userID, keyID string) error
-	UpdateAccountPeers(ctx context.Context, accountID string)
-	BufferUpdateAccountPeers(ctx context.Context, accountID string)
+	UpdateAccountPeers(ctx context.Context, accountID string, reason types.UpdateReason)
+	ExpandAndUpdateAffected(ctx context.Context, accountID string, snap *affectedpeers.Snapshot, change affectedpeers.Change)
+	BufferUpdateAccountPeers(ctx context.Context, accountID string, reason types.UpdateReason)
 	BuildUserInfosForAccount(ctx context.Context, accountID, initiatorUserID string, accountUsers []*types.User) (map[string]*types.UserInfo, error)
 	SyncUserJWTGroups(ctx context.Context, userAuth auth.UserAuth) error
 	GetStore() store.Store

management/server/account/pat.go

@@ -0,0 +1,8 @@
+package account
+
+const (
+	// PATMinExpireDays is the minimum allowed Personal Access Token expiration in days.
+	PATMinExpireDays = 1
+	// PATMaxExpireDays is the maximum allowed Personal Access Token expiration in days.
+	PATMaxExpireDays = 365
+)

management/server/account/request_buffer.go

@@ -6,6 +6,8 @@ import (
 	"github.com/netbirdio/netbird/management/server/types"
 )
 
+//go:generate go tool mockgen -package=account -source=./request_buffer.go -destination=request_buffer_mock.go
+
 type RequestBuffer interface {
 	GetAccountWithBackpressure(ctx context.Context, accountID string) (*types.Account, error)
 }

management/server/account/request_buffer_mock.go

@@ -0,0 +1,57 @@
+// Code generated by MockGen. DO NOT EDIT.
+// Source: ./request_buffer.go
+//
+// Generated by this command:
+//
+//	mockgen -package=account -source=./request_buffer.go -destination=request_buffer_mock.go
+//
+
+// Package account is a generated GoMock package.
+package account
+
+import (
+	context "context"
+	reflect "reflect"
+
+	types "github.com/netbirdio/netbird/management/server/types"
+	gomock "go.uber.org/mock/gomock"
+)
+
+// MockRequestBuffer is a mock of RequestBuffer interface.
+type MockRequestBuffer struct {
+	ctrl     *gomock.Controller
+	recorder *MockRequestBufferMockRecorder
+	isgomock struct{}
+}
+
+// MockRequestBufferMockRecorder is the mock recorder for MockRequestBuffer.
+type MockRequestBufferMockRecorder struct {
+	mock *MockRequestBuffer
+}
+
+// NewMockRequestBuffer creates a new mock instance.
+func NewMockRequestBuffer(ctrl *gomock.Controller) *MockRequestBuffer {
+	mock := &MockRequestBuffer{ctrl: ctrl}
+	mock.recorder = &MockRequestBufferMockRecorder{mock}
+	return mock
+}
+
+// EXPECT returns an object that allows the caller to indicate expected use.
+func (m *MockRequestBuffer) EXPECT() *MockRequestBufferMockRecorder {
+	return m.recorder
+}
+
+// GetAccountWithBackpressure mocks base method.
+func (m *MockRequestBuffer) GetAccountWithBackpressure(ctx context.Context, accountID string) (*types.Account, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountWithBackpressure", ctx, accountID)
+	ret0, _ := ret[0].(*types.Account)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountWithBackpressure indicates an expected call of GetAccountWithBackpressure.
+func (mr *MockRequestBufferMockRecorder) GetAccountWithBackpressure(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountWithBackpressure", reflect.TypeOf((*MockRequestBuffer)(nil).GetAccountWithBackpressure), ctx, accountID)
+}

management/server/account_request_buffer.go

@@ -2,117 +2,38 @@ package server
 
 import (
 	"context"
-	"os"
-	"sync"
 	"time"
 
 	log "github.com/sirupsen/logrus"
 
+	"github.com/netbirdio/netbird/management/internals/shared/requestbuffer"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 )
 
-// AccountRequest holds the result channel to return the requested account.
-type AccountRequest struct {
-	AccountID  string
-	ResultChan chan *AccountResult
-}
-
-// AccountResult holds the account data or an error.
-type AccountResult struct {
-	Account *types.Account
-	Err     error
-}
+const defaultAccountBufferInterval = 100 * time.Millisecond
 
 type AccountRequestBuffer struct {
-	store               store.Store
-	getAccountRequests  map[string][]*AccountRequest
-	mu                  sync.Mutex
-	getAccountRequestCh chan *AccountRequest
-	bufferInterval      time.Duration
+	buffer *requestbuffer.Buffer[*types.Account]
 }
 
 func NewAccountRequestBuffer(ctx context.Context, store store.Store) *AccountRequestBuffer {
-	bufferIntervalStr := os.Getenv("NB_GET_ACCOUNT_BUFFER_INTERVAL")
-	bufferInterval, err := time.ParseDuration(bufferIntervalStr)
-	if err != nil {
-		if bufferIntervalStr != "" {
-			log.WithContext(ctx).Warnf("failed to parse account request buffer interval: %s", err)
-		}
-		bufferInterval = 100 * time.Millisecond
+	interval := requestbuffer.Interval(ctx, "NB_GET_ACCOUNT_BUFFER_INTERVAL", defaultAccountBufferInterval)
+	log.WithContext(ctx).Infof("set account request buffer interval to %s", interval)
+
+	return &AccountRequestBuffer{
+		buffer: requestbuffer.New(ctx, "account request buffer", interval, store.GetAccount),
 	}
-
-	log.WithContext(ctx).Infof("set account request buffer interval to %s", bufferInterval)
-
-	ac := AccountRequestBuffer{
-		store:               store,
-		getAccountRequests:  make(map[string][]*AccountRequest),
-		getAccountRequestCh: make(chan *AccountRequest),
-		bufferInterval:      bufferInterval,
-	}
-
-	go ac.processGetAccountRequests(ctx)
-
-	return &ac
 }
+
 func (ac *AccountRequestBuffer) GetAccountWithBackpressure(ctx context.Context, accountID string) (*types.Account, error) {
-	req := &AccountRequest{
-		AccountID:  accountID,
-		ResultChan: make(chan *AccountResult, 1),
+	account, err := ac.buffer.Get(ctx, accountID)
+	if err != nil || account == nil {
+		return account, err
 	}
 
-	log.WithContext(ctx).Tracef("requesting account %s with backpressure", accountID)
-	startTime := time.Now()
-	ac.getAccountRequestCh <- req
-
-	result := <-req.ResultChan
-	log.WithContext(ctx).Tracef("got account with backpressure after %s", time.Since(startTime))
-	return result.Account, result.Err
-}
-
-func (ac *AccountRequestBuffer) processGetAccountBatch(ctx context.Context, accountID string) {
-	ac.mu.Lock()
-	requests := ac.getAccountRequests[accountID]
-	delete(ac.getAccountRequests, accountID)
-	ac.mu.Unlock()
-
-	if len(requests) == 0 {
-		return
-	}
-
-	startTime := time.Now()
-	account, err := ac.store.GetAccount(ctx, accountID)
-	log.WithContext(ctx).Tracef("getting account %s in batch took %s", accountID, time.Since(startTime))
-	result := &AccountResult{Account: account, Err: err}
-
-	for _, req := range requests {
-		if account != nil {
-			// Shallow copy the account so each goroutine gets its own struct value.
-			// This prevents data races when callers mutate fields like Policies.
-			accountCopy := *account
-			req.ResultChan <- &AccountResult{Account: &accountCopy, Err: err}
-		} else {
-			req.ResultChan <- result
-		}
-		close(req.ResultChan)
-	}
-}
-
-func (ac *AccountRequestBuffer) processGetAccountRequests(ctx context.Context) {
-	for {
-		select {
-		case req := <-ac.getAccountRequestCh:
-			ac.mu.Lock()
-			ac.getAccountRequests[req.AccountID] = append(ac.getAccountRequests[req.AccountID], req)
-			if len(ac.getAccountRequests[req.AccountID]) == 1 {
-				go func(ctx context.Context, accountID string) {
-					time.Sleep(ac.bufferInterval)
-					ac.processGetAccountBatch(ctx, accountID)
-				}(ctx, req.AccountID)
-			}
-			ac.mu.Unlock()
-		case <-ctx.Done():
-			return
-		}
-	}
+	// Shallow copy the account so each caller gets its own struct value.
+	// This prevents data races when callers mutate fields like Policies.
+	accountCopy := *account
+	return &accountCopy, nil
 }

management/server/activity/codes.go

@@ -231,6 +231,58 @@ const (
 	DomainDeleted Activity = 119
 	// DomainValidated indicates that a custom domain was validated
 	DomainValidated Activity = 120
+	// AccountIPv6Enabled indicates that a user enabled IPv6 overlay for the account
+	AccountIPv6Enabled Activity = 121
+	// AccountIPv6Disabled indicates that a user disabled IPv6 overlay for the account
+	AccountIPv6Disabled Activity = 122
+
+	// AccountLocalMfaEnabled indicates that a user enabled TOTP MFA for local users
+	AccountLocalMfaEnabled Activity = 123
+	// AccountLocalMfaDisabled indicates that a user disabled TOTP MFA for local users
+	AccountLocalMfaDisabled Activity = 124
+	// UserExtendedPeerSession indicates that a user refreshed their peer's
+	// SSO session deadline via ExtendAuthSession without re-establishing the
+	// tunnel. Distinct from UserLoggedInPeer (full interactive login).
+	UserExtendedPeerSession Activity = 125
+
+	// AgentNetworkProviderCreated indicates that a user created an Agent Network provider
+	AgentNetworkProviderCreated Activity = 126
+	// AgentNetworkProviderUpdated indicates that a user updated an Agent Network provider
+	AgentNetworkProviderUpdated Activity = 127
+	// AgentNetworkProviderDeleted indicates that a user deleted an Agent Network provider
+	AgentNetworkProviderDeleted Activity = 128
+
+	// AgentNetworkPolicyCreated indicates that a user created an Agent Network policy
+	AgentNetworkPolicyCreated Activity = 129
+	// AgentNetworkPolicyUpdated indicates that a user updated an Agent Network policy
+	AgentNetworkPolicyUpdated Activity = 130
+	// AgentNetworkPolicyDeleted indicates that a user deleted an Agent Network policy
+	AgentNetworkPolicyDeleted Activity = 131
+
+	// AgentNetworkGuardrailCreated indicates that a user created an Agent Network guardrail
+	AgentNetworkGuardrailCreated Activity = 132
+	// AgentNetworkGuardrailUpdated indicates that a user updated an Agent Network guardrail
+	AgentNetworkGuardrailUpdated Activity = 133
+	// AgentNetworkGuardrailDeleted indicates that a user deleted an Agent Network guardrail
+	AgentNetworkGuardrailDeleted Activity = 134
+
+	// AgentNetworkBudgetRuleCreated indicates that a user created an Agent Network budget rule
+	AgentNetworkBudgetRuleCreated Activity = 135
+	// AgentNetworkBudgetRuleUpdated indicates that a user updated an Agent Network budget rule
+	AgentNetworkBudgetRuleUpdated Activity = 136
+	// AgentNetworkBudgetRuleDeleted indicates that a user deleted an Agent Network budget rule
+	AgentNetworkBudgetRuleDeleted Activity = 137
+
+	// AgentNetworkSettingsUpdated indicates that a user updated Agent Network account settings
+	AgentNetworkSettingsUpdated Activity = 139
+
+	// AccountMetricsPushEnabled indicates that a user enabled metrics push for the account
+	AccountMetricsPushEnabled Activity = 140
+	// AccountMetricsPushDisabled indicates that a user disabled metrics push for the account
+	AccountMetricsPushDisabled Activity = 141
+
+	// AgentNetworkSettingsDeleted indicates that a user deleted the Agent Network account settings, releasing the endpoint
+	AgentNetworkSettingsDeleted Activity = 142
 
 	AccountDeleted Activity = 99999
 )
@@ -347,6 +399,9 @@ var activityMap = map[Activity]Code{
 	AccountAutoUpdateAlwaysEnabled:  {"Account auto-update always enabled", "account.setting.auto.update.always.enable"},
 	AccountAutoUpdateAlwaysDisabled: {"Account auto-update always disabled", "account.setting.auto.update.always.disable"},
 
+	AccountIPv6Enabled:  {"Account IPv6 overlay enabled", "account.setting.ipv6.enable"},
+	AccountIPv6Disabled: {"Account IPv6 overlay disabled", "account.setting.ipv6.disable"},
+
 	IdentityProviderCreated: {"Identity provider created", "identityprovider.create"},
 	IdentityProviderUpdated: {"Identity provider updated", "identityprovider.update"},
 	IdentityProviderDeleted: {"Identity provider deleted", "identityprovider.delete"},
@@ -379,6 +434,33 @@ var activityMap = map[Activity]Code{
 	AccountPeerExposeEnabled:  {"Account peer expose enabled", "account.setting.peer.expose.enable"},
 	AccountPeerExposeDisabled: {"Account peer expose disabled", "account.setting.peer.expose.disable"},
 
+	AccountLocalMfaEnabled:  {"Account local MFA enabled", "account.setting.local.mfa.enable"},
+	AccountLocalMfaDisabled: {"Account local MFA disabled", "account.setting.local.mfa.disable"},
+
+	UserExtendedPeerSession: {"User extended peer session", "user.peer.session.extend"},
+
+	AgentNetworkProviderCreated: {"Agent Network provider created", "agent_network.provider.create"},
+	AgentNetworkProviderUpdated: {"Agent Network provider updated", "agent_network.provider.update"},
+	AgentNetworkProviderDeleted: {"Agent Network provider deleted", "agent_network.provider.delete"},
+
+	AgentNetworkPolicyCreated: {"Agent Network policy created", "agent_network.policy.create"},
+	AgentNetworkPolicyUpdated: {"Agent Network policy updated", "agent_network.policy.update"},
+	AgentNetworkPolicyDeleted: {"Agent Network policy deleted", "agent_network.policy.delete"},
+
+	AgentNetworkGuardrailCreated: {"Agent Network guardrail created", "agent_network.guardrail.create"},
+	AgentNetworkGuardrailUpdated: {"Agent Network guardrail updated", "agent_network.guardrail.update"},
+	AgentNetworkGuardrailDeleted: {"Agent Network guardrail deleted", "agent_network.guardrail.delete"},
+
+	AgentNetworkBudgetRuleCreated: {"Agent Network budget rule created", "agent_network.budget_rule.create"},
+	AgentNetworkBudgetRuleUpdated: {"Agent Network budget rule updated", "agent_network.budget_rule.update"},
+	AgentNetworkBudgetRuleDeleted: {"Agent Network budget rule deleted", "agent_network.budget_rule.delete"},
+
+	AgentNetworkSettingsUpdated: {"Agent Network settings updated", "agent_network.settings.update"},
+	AgentNetworkSettingsDeleted: {"Agent Network settings deleted", "agent_network.settings.delete"},
+
+	AccountMetricsPushEnabled:  {"Account metrics push enabled", "account.setting.metrics.push.enable"},
+	AccountMetricsPushDisabled: {"Account metrics push disabled", "account.setting.metrics.push.disable"},
+
 	DomainAdded:     {"Domain added", "domain.add"},
 	DomainDeleted:   {"Domain deleted", "domain.delete"},
 	DomainValidated: {"Domain validated", "domain.validate"},

management/server/affected_peers_coverage_test.go

@@ -0,0 +1,115 @@
+package server
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
+	resourceTypes "github.com/netbirdio/netbird/management/server/networks/resources/types"
+	networkTypes "github.com/netbirdio/netbird/management/server/networks/types"
+	"github.com/netbirdio/netbird/management/server/posture"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// TestAffectedPeers_DependencyCoverageMatrix enumerates each network-map
+// dependency crossed with the change-type that can alter it, asserting the
+// resolver folds in exactly the peers whose map changes. A new dependency that
+// the resolver fails to walk should fail one of these rows; a new change-type
+// without a row is a coverage gap to add here.
+func TestAffectedPeers_DependencyCoverageMatrix(t *testing.T) {
+	type row struct {
+		name  string
+		build func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string)
+	}
+
+	rows := []row{
+		{
+			name: "policy-groups/source-group-change refreshes source+routing, excludes unrelated",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+				require.NoError(t, err)
+				return affectedpeers.Change{ChangedGroupIDs: []string{s.sourceGroupID}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+		{
+			name: "resource-routing-bridge/router-peer-change refreshes policy sources",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+				require.NoError(t, err)
+				return affectedpeers.Change{ChangedPeerIDs: []string{s.routerPeerID}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+		{
+			name: "policy-change/explicit-policy refreshes source+routing",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+				return affectedpeers.Change{Policies: []*types.Policy{policy}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+		{
+			name: "policy-destinationresource/explicit-policy bridges to routing peer",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				policy := peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID)
+				return affectedpeers.Change{Policies: []*types.Policy{policy}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+		{
+			name: "resource-change refreshes source+routing on its network",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+				require.NoError(t, err)
+				return affectedpeers.Change{Resources: []*resourceTypes.NetworkResource{
+						{ID: s.resourceID, NetworkID: s.networkID, GroupIDs: []string{s.resourceGroupID}},
+					}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+		{
+			name: "network-change refreshes source+routing on that network",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+				require.NoError(t, err)
+				return affectedpeers.Change{Networks: []*networkTypes.Network{{ID: s.networkID}}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+		{
+			name: "posture-check-change refreshes source+routing of gated policy",
+			build: func(t *testing.T, s *routerScenario, ctx context.Context) (affectedpeers.Change, []string, []string) {
+				check, err := s.manager.SavePostureChecks(ctx, s.accountID, userID, &posture.Checks{
+					Name:   "cov-min-version",
+					Checks: posture.ChecksDefinition{NBVersionCheck: &posture.NBVersionCheck{MinVersion: "0.30.0"}},
+				}, true)
+				require.NoError(t, err)
+				policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+				policy.SourcePostureChecks = []string{check.ID}
+				_, err = s.manager.SavePolicy(ctx, s.accountID, userID, policy, true)
+				require.NoError(t, err)
+				return affectedpeers.Change{PostureCheckIDs: []string{check.ID}},
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
+			},
+		},
+	}
+
+	for _, r := range rows {
+		t.Run(r.name, func(t *testing.T) {
+			s := setupRouterScenario(t, true)
+			ctx := context.Background()
+
+			change, mustContain, mustExclude := r.build(t, s, ctx)
+			affected := resolveAffected(t, s.manager.Store, s.accountID, change)
+
+			assert.ElementsMatch(t, affected, mustContain, "expected peer to be affected")
+			for _, peerID := range mustExclude {
+				assert.NotContains(t, affected, peerID, "peer must not be affected")
+			}
+		})
+	}
+}

management/server/affected_peers_jwt_test.go

@@ -0,0 +1,179 @@
+package server
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/auth"
+)
+
+// A user's auto-group change refreshes the destinations of the SSH rules authorizing
+// that group — they carry the group -> user mapping — even though no peer moved
+// between groups.
+func TestAffectedPeers_UserGroupChange_RefreshesSSHAuthorizedDestinations(t *testing.T) {
+	manager, s, accountID, peerIDs, groupIDs := setupAffectedPeersTest(t)
+	ctx := context.Background()
+
+	_, err := manager.SavePolicy(ctx, accountID, userID, &types.Policy{
+		Enabled: true,
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:          true,
+				Sources:          []string{groupIDs[0]},
+				Destinations:     []string{groupIDs[1]},
+				Protocol:         types.PolicyRuleProtocolNetbirdSSH,
+				Action:           types.PolicyTrafficActionAccept,
+				AuthorizedGroups: map[string][]string{groupIDs[3]: {"root"}},
+			},
+		},
+	}, true)
+	require.NoError(t, err)
+
+	result := resolveAffected(t, s, accountID, affectedpeers.Change{UserGroupIDs: []string{groupIDs[3]}})
+	assert.ElementsMatch(t, []string{peerIDs[1]}, result,
+		"only the SSH rule's destination peers carry the changed group -> user mapping")
+
+	result = resolveAffected(t, s, accountID, affectedpeers.Change{UserGroupIDs: []string{groupIDs[4]}})
+	assert.Empty(t, result, "a group no SSH rule authorizes affects nobody")
+}
+
+// Creating, blocking or unblocking a user changes the account's allowed-user set, which
+// reaches only the destinations of the SSH rules that ship it.
+func TestAffectedPeers_AllowedUsersChange_RefreshesSSHDestinations(t *testing.T) {
+	manager, s, accountID, peerIDs, groupIDs := setupAffectedPeersTest(t)
+	ctx := context.Background()
+
+	// Ships the allowed-user set: an SSH rule naming no groups and no user.
+	_, err := manager.SavePolicy(ctx, accountID, userID, &types.Policy{
+		Enabled: true,
+		Rules: []*types.PolicyRule{{
+			Enabled:      true,
+			Sources:      []string{groupIDs[0]},
+			Destinations: []string{groupIDs[1]},
+			Protocol:     types.PolicyRuleProtocolNetbirdSSH,
+			Action:       types.PolicyTrafficActionAccept,
+		}},
+	}, true)
+	require.NoError(t, err)
+
+	// Does not ship it: an SSH rule that authorizes a specific group.
+	_, err = manager.SavePolicy(ctx, accountID, userID, &types.Policy{
+		Enabled: true,
+		Rules: []*types.PolicyRule{{
+			Enabled:          true,
+			Sources:          []string{groupIDs[2]},
+			Destinations:     []string{groupIDs[3]},
+			Protocol:         types.PolicyRuleProtocolNetbirdSSH,
+			Action:           types.PolicyTrafficActionAccept,
+			AuthorizedGroups: map[string][]string{groupIDs[0]: {"root"}},
+		}},
+	}, true)
+	require.NoError(t, err)
+
+	result := resolveAffected(t, s, accountID, affectedpeers.Change{AllowedUsersChanged: true})
+	assert.ElementsMatch(t, []string{peerIDs[1]}, result,
+		"only the destinations of the rule shipping the allowed-user set refresh")
+}
+
+// TestAffectedPeers_SyncUserJWTGroups_OnlyAffectedPeersUpdated verifies that a JWT
+// auto-group change updates only the user's peers and the peers linked to the changed
+// group through policies, instead of fanning out to the whole account.
+func TestAffectedPeers_SyncUserJWTGroups_OnlyAffectedPeersUpdated(t *testing.T) {
+	manager, updateManager, account, _, peer2, peer3 := setupNetworkMapTest(t)
+	ctx := context.Background()
+	accountID := account.Id
+
+	key, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err)
+	userPeer, _, _, _, err := manager.AddPeer(ctx, accountID, "", userID, &nbpeer.Peer{
+		Key:  key.PublicKey().String(),
+		Meta: nbpeer.PeerSystemMeta{Hostname: "user-peer"},
+	}, false)
+	require.NoError(t, err)
+
+	policies, err := manager.Store.GetAccountPolicies(ctx, store.LockingStrengthNone, accountID)
+	require.NoError(t, err)
+	for _, p := range policies {
+		require.NoError(t, manager.Store.DeletePolicy(ctx, accountID, p.ID))
+	}
+
+	account, err = manager.Store.GetAccount(ctx, accountID)
+	require.NoError(t, err)
+	account.Settings.JWTGroupsEnabled = true
+	account.Settings.JWTGroupsClaimName = "groups"
+	account.Settings.GroupsPropagationEnabled = true
+	require.NoError(t, manager.Store.SaveAccount(ctx, account))
+
+	require.NoError(t, manager.CreateGroup(ctx, accountID, userID, &types.Group{ID: "jwt-grp", Name: "jwt-linked", Issued: types.GroupIssuedJWT, Peers: []string{}}))
+	require.NoError(t, manager.CreateGroup(ctx, accountID, userID, &types.Group{ID: "jwt-dest", Name: "jwt-dest", Peers: []string{peer2.ID}}))
+
+	_, err = manager.SavePolicy(ctx, accountID, userID, &types.Policy{
+		Enabled: true,
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:       true,
+				Sources:       []string{"jwt-grp"},
+				Destinations:  []string{"jwt-dest"},
+				Bidirectional: true,
+				Action:        types.PolicyTrafficActionAccept,
+			},
+		},
+	}, true)
+	require.NoError(t, err)
+
+	updUser := updateManager.CreateChannel(ctx, userPeer.ID)
+	upd2 := updateManager.CreateChannel(ctx, peer2.ID)
+	upd3 := updateManager.CreateChannel(ctx, peer3.ID)
+	t.Cleanup(func() {
+		updateManager.CloseChannel(ctx, userPeer.ID)
+		updateManager.CloseChannel(ctx, peer2.ID)
+		updateManager.CloseChannel(ctx, peer3.ID)
+	})
+
+	userAuth := auth.UserAuth{
+		AccountId: accountID,
+		UserId:    userID,
+		Groups:    []string{"jwt-linked"},
+	}
+
+	t.Run("adding JWT group updates only linked peers", func(t *testing.T) {
+		drainPeerUpdates(updUser)
+		drainPeerUpdates(upd2)
+		drainPeerUpdates(upd3)
+
+		require.NoError(t, manager.SyncUserJWTGroups(ctx, userAuth))
+
+		peerShouldReceiveUpdate(t, updUser)
+		peerShouldReceiveUpdate(t, upd2)
+		peerShouldNotReceiveUpdate(t, upd3)
+
+		user, err := manager.Store.GetUserByUserID(ctx, store.LockingStrengthNone, userID)
+		require.NoError(t, err)
+		assert.Contains(t, user.AutoGroups, "jwt-grp")
+	})
+
+	t.Run("removing JWT group updates only linked peers", func(t *testing.T) {
+		drainPeerUpdates(updUser)
+		drainPeerUpdates(upd2)
+		drainPeerUpdates(upd3)
+
+		userAuth.Groups = nil
+		require.NoError(t, manager.SyncUserJWTGroups(ctx, userAuth))
+
+		peerShouldReceiveUpdate(t, updUser)
+		peerShouldReceiveUpdate(t, upd2)
+		peerShouldNotReceiveUpdate(t, upd3)
+
+		user, err := manager.Store.GetUserByUserID(ctx, store.LockingStrengthNone, userID)
+		require.NoError(t, err)
+		assert.NotContains(t, user.AutoGroups, "jwt-grp")
+	})
+}

management/server/affected_peers_oldstate_test.go

@@ -0,0 +1,143 @@
+package server
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+
+	resourceTypes "github.com/netbirdio/netbird/management/server/networks/resources/types"
+	routerTypes "github.com/netbirdio/netbird/management/server/networks/routers/types"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// An update spans an old and a new state. The affected set must be the UNION of
+// peers reachable before and after the change; resolving only against the final
+// state drops peers that were reachable but no longer are. These tests pin the
+// two paths where the old state is reachable only by the changed object's
+// previous references: detaching a resource group, and re-pointing a router peer.
+
+// TestAffectedPeers_E2E_UpdateResource_DetachGroup_RefreshesOldGroupSources:
+// a resource is reachable by a source group via two destination resource groups;
+// detaching one of them must still refresh that group's policy source peers, even
+// though the post-update resource no longer maps to it.
+func TestAffectedPeers_E2E_UpdateResource_DetachGroup_RefreshesOldGroupSources(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	// A second resource group + a second source group/peer that reaches the
+	// resource only through that second group.
+	const detachGroupID = "rs-detach-grp"
+	require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{ID: detachGroupID, Name: "rs-detach"}))
+
+	const secondSourceGroupID = "rs-source-grp-2"
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "rs-detach-key", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	secondSourcePeer := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+	require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{
+		ID: secondSourceGroupID, Name: "rs-source-2", Peers: []string{secondSourcePeer.ID},
+	}))
+
+	resourcesManager, _, _ := s.managers()
+
+	// Attach the resource to the detach group as well: now in [resourceGroup, detachGroup].
+	_, err = resourcesManager.UpdateResource(ctx, userID, &resourceTypes.NetworkResource{
+		ID:        s.resourceID,
+		AccountID: s.accountID,
+		NetworkID: s.networkID,
+		Name:      "rs-resource-host",
+		Address:   "10.20.30.0/24",
+		GroupIDs:  []string{s.resourceGroupID, detachGroupID},
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	// Policy granting the second source group access via the detach group.
+	_, err = s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(secondSourceGroupID, detachGroupID), true)
+	require.NoError(t, err)
+
+	secondSrcCh := s.updateManager.CreateChannel(ctx, secondSourcePeer.ID)
+	t.Cleanup(func() { s.updateManager.CloseChannel(ctx, secondSourcePeer.ID) })
+	settleAffectedUpdates(secondSrcCh)
+
+	done := make(chan struct{})
+	go func() {
+		// Detaching the resource from detachGroup removes the second source's
+		// access; that source peer must be refreshed even though the post-update
+		// resource no longer maps to detachGroup.
+		peerShouldReceiveUpdate(t, secondSrcCh)
+		close(done)
+	}()
+
+	_, err = resourcesManager.UpdateResource(ctx, userID, &resourceTypes.NetworkResource{
+		ID:        s.resourceID,
+		AccountID: s.accountID,
+		NetworkID: s.networkID,
+		Name:      "rs-resource-host",
+		Address:   "10.20.30.0/24",
+		GroupIDs:  []string{s.resourceGroupID}, // detached detachGroup
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: detaching a resource group did not refresh the old group's policy source peer")
+	}
+}
+
+// TestAffectedPeers_E2E_UpdateRouter_RepointPeer_RefreshesOldRoutingPeer:
+// changing router.Peer within the same network must still refresh the OLD routing
+// peer, which loses its routing role.
+func TestAffectedPeers_E2E_UpdateRouter_RepointPeer_RefreshesOldRoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	_, routersManager, _ := s.managers()
+
+	routers, err := s.manager.Store.GetNetworkRoutersByNetID(ctx, store.LockingStrengthNone, s.accountID, s.networkID)
+	require.NoError(t, err)
+	require.Len(t, routers, 1)
+	router := routers[0]
+	oldRoutingPeer := router.Peer
+	require.NotEmpty(t, oldRoutingPeer)
+
+	// A new peer to become the routing peer in place of the old one.
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "rs-newrouter-key", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	newRoutingPeer := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+
+	oldCh := s.updateManager.CreateChannel(ctx, oldRoutingPeer)
+	t.Cleanup(func() { s.updateManager.CloseChannel(ctx, oldRoutingPeer) })
+	settleAffectedUpdates(oldCh)
+
+	done := make(chan struct{})
+	go func() {
+		// The old routing peer stops serving the resource and must be refreshed.
+		peerShouldReceiveUpdate(t, oldCh)
+		close(done)
+	}()
+
+	_, err = routersManager.UpdateRouter(ctx, userID, &routerTypes.NetworkRouter{
+		ID:         router.ID,
+		NetworkID:  s.networkID,
+		AccountID:  s.accountID,
+		Peer:       newRoutingPeer.ID, // repoint within the same network
+		Masquerade: true,
+		Metric:     9999,
+		Enabled:    true,
+	})
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: re-pointing the router peer did not refresh the old routing peer")
+	}
+}

management/server/affected_peers_property_test.go

@@ -0,0 +1,253 @@
+package server
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"math/rand"
+	"sort"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+	"golang.org/x/exp/maps"
+
+	nbdns "github.com/netbirdio/netbird/dns"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// allPeerMaps computes the serialized per-peer network map for every peer in the
+// account, mirroring the controller's compute path so the property test compares
+// against real output.
+func allPeerMaps(t *testing.T, manager *DefaultAccountManager, accountID string) map[string]string {
+	t.Helper()
+	ctx := context.Background()
+
+	account, err := manager.Store.GetAccount(ctx, accountID)
+	require.NoError(t, err)
+
+	validated := make(map[string]struct{}, len(account.Peers))
+	for id := range account.Peers {
+		validated[id] = struct{}{}
+	}
+	resourcePolicies := account.GetResourcePoliciesMap()
+	routers := account.GetResourceRoutersMap()
+	groupIDToUserIDs := account.GetActiveGroupUsers()
+
+	out := make(map[string]string, len(account.Peers))
+	for peerID := range account.Peers {
+		nm := account.GetPeerNetworkMapFromComponents(ctx, peerID, nbdns.CustomZone{}, nil, validated, resourcePolicies, routers, nil, groupIDToUserIDs)
+		// Network.Serial is an account-global counter bumped on every change; it
+		// is not a per-peer dependency, so normalize it out of the comparison.
+		if nm.Network != nil {
+			nm.Network.Serial = 0
+		}
+		out[peerID] = canonicalJSON(t, nm)
+	}
+	return out
+}
+
+// canonicalJSON marshals v and returns an order-insensitive string form: every
+// JSON array is sorted by the canonical form of its elements. The network map's
+// Peers/Routes/FirewallRules/SourceRanges slices have nondeterministic order, so
+// a raw JSON compare would report spurious changes.
+func canonicalJSON(t *testing.T, v interface{}) string {
+	t.Helper()
+	b, err := json.Marshal(v)
+	require.NoError(t, err)
+	var parsed interface{}
+	require.NoError(t, json.Unmarshal(b, &parsed))
+	canonicalized, err := json.Marshal(sortAny(parsed))
+	require.NoError(t, err)
+	return string(canonicalized)
+}
+
+func sortAny(v interface{}) interface{} {
+	switch val := v.(type) {
+	case []interface{}:
+		for i := range val {
+			val[i] = sortAny(val[i])
+		}
+		sort.Slice(val, func(i, j int) bool {
+			bi, _ := json.Marshal(val[i])
+			bj, _ := json.Marshal(val[j])
+			return string(bi) < string(bj)
+		})
+		return val
+	case map[string]interface{}:
+		for k := range val {
+			val[k] = sortAny(val[k])
+		}
+		return val
+	default:
+		return v
+	}
+}
+
+// changedPeers returns the peer IDs whose serialized map differs between before
+// and after.
+func changedPeers(before, after map[string]string) []string {
+	var changed []string
+	for id, b := range before {
+		a, ok := after[id]
+		if !ok || a != b {
+			changed = append(changed, id)
+		}
+	}
+	for id := range after {
+		if _, ok := before[id]; !ok {
+			changed = append(changed, id)
+		}
+	}
+	return changed
+}
+
+// TestAffectedPeers_Property_ResolverSupersetsRealChanges builds a topology,
+// applies random changes, and asserts that the resolver's affected set is a
+// superset of the peers whose real network map actually changed. If the resolver
+// ever misses a dependency, a change will alter a peer's map without that peer
+// appearing in the affected set, failing here.
+func TestAffectedPeers_Property_ResolverSupersetsRealChanges(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	// A pre-existing peer->resource policy so the resource/router bridge is live.
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	// Extra peers and groups to give mutations room to move membership around.
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "prop-key", types.SetupKeyReusable, 0, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	extraPeers := make([]string, 0, 4)
+	for i := 0; i < 4; i++ {
+		p := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+		extraPeers = append(extraPeers, p.ID)
+	}
+	extraGroups := []string{"prop-grp-0", "prop-grp-1"}
+	for _, g := range extraGroups {
+		require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{ID: g, Name: g}))
+	}
+
+	rng := rand.New(rand.NewSource(1))
+	allGroups := append([]string{s.sourceGroupID, s.resourceGroupID, s.routerPeerGroupID}, extraGroups...)
+	allPeers := append([]string{s.sourcePeerID, s.routerPeerID, s.routerGroupPeerID, s.unrelatedPeerID}, extraPeers...)
+
+	for iter := 0; iter < 60; iter++ {
+		change, apply := s.randomMutation(t, rng, allGroups, allPeers)
+		if apply == nil {
+			continue
+		}
+
+		before := allPeerMaps(t, s.manager, s.accountID)
+
+		resolvedSet := make(map[string]struct{})
+		resolve := func() {
+			require.NoError(t, s.manager.Store.ExecuteInTransaction(ctx, func(tx store.Store) error {
+				snap, err := affectedpeers.Load(ctx, tx, s.accountID, change)
+				if err != nil {
+					return err
+				}
+				for _, id := range snap.Expand(ctx, s.accountID, change) {
+					resolvedSet[id] = struct{}{}
+				}
+				return nil
+			}))
+		}
+
+		// Resolve on both sides of the mutation and union: removals are visible
+		// only pre-apply (the leaving peer is still a member), additions only
+		// post-apply (the joining peer is now a member). Production captures both
+		// via per-path handling (e.g. UpdateGroup passes peersToRemove); the union
+		// models that without coupling the test to each path's ordering.
+		resolve()
+		changedIDs := change.ChangedPeerIDs
+		apply()
+		resolve()
+
+		after := allPeerMaps(t, s.manager, s.accountID)
+
+		// The explicitly-changed peer's own map refresh is the caller's
+		// responsibility (the resolver returns the peers to propagate to), so it
+		// is allowed to be absent from the resolved set.
+		changedExplicitly := make(map[string]struct{}, len(changedIDs))
+		for _, id := range changedIDs {
+			changedExplicitly[id] = struct{}{}
+		}
+
+		for _, id := range changedPeers(before, after) {
+			if _, stillExists := after[id]; !stillExists {
+				continue
+			}
+			if _, isExplicit := changedExplicitly[id]; isExplicit {
+				continue
+			}
+			_, ok := resolvedSet[id]
+			require.Truef(t, ok,
+				"iter %d: peer %s network map changed but was not in the resolver's affected set %v (change=%+v)",
+				iter, id, maps.Keys(resolvedSet), change)
+		}
+	}
+}
+
+// randomMutation picks a random change, returns the Change to resolve and a
+// function that applies the underlying store mutation. apply is nil when the
+// drawn mutation is a no-op for the current state.
+func (s *routerScenario) randomMutation(t *testing.T, rng *rand.Rand, allGroups, allPeers []string) (affectedpeers.Change, func()) {
+	t.Helper()
+	ctx := context.Background()
+
+	switch rng.Intn(3) {
+	case 0:
+		groupID := allGroups[rng.Intn(len(allGroups))]
+		peerID := allPeers[rng.Intn(len(allPeers))]
+		grp, err := s.manager.Store.GetGroupByID(ctx, store.LockingStrengthNone, s.accountID, groupID)
+		require.NoError(t, err)
+		if slicesContains(grp.Peers, peerID) {
+			return affectedpeers.Change{}, nil
+		}
+		return affectedpeers.Change{ChangedGroupIDs: []string{groupID}, ChangedPeerIDs: []string{peerID}},
+			func() {
+				require.NoError(t, s.manager.GroupAddPeer(ctx, s.accountID, groupID, peerID))
+			}
+	case 1:
+		groupID := allGroups[rng.Intn(len(allGroups))]
+		grp, err := s.manager.Store.GetGroupByID(ctx, store.LockingStrengthNone, s.accountID, groupID)
+		require.NoError(t, err)
+		if len(grp.Peers) == 0 {
+			return affectedpeers.Change{}, nil
+		}
+		peerID := grp.Peers[rng.Intn(len(grp.Peers))]
+		return affectedpeers.Change{ChangedGroupIDs: []string{groupID}, ChangedPeerIDs: []string{peerID}},
+			func() {
+				require.NoError(t, s.manager.GroupDeletePeer(ctx, s.accountID, groupID, peerID))
+			}
+	default:
+		src := allGroups[rng.Intn(len(allGroups))]
+		dst := allGroups[rng.Intn(len(allGroups))]
+		policy := &types.Policy{
+			Enabled: true,
+			Name:    fmt.Sprintf("prop-policy-%d", rng.Int()),
+			Rules: []*types.PolicyRule{{
+				Enabled:      true,
+				Sources:      []string{src},
+				Destinations: []string{dst},
+				Action:       types.PolicyTrafficActionAccept,
+			}},
+		}
+		return affectedpeers.Change{Policies: []*types.Policy{policy}},
+			func() {
+				_, err := s.manager.SavePolicy(ctx, s.accountID, userID, policy, true)
+				require.NoError(t, err)
+			}
+	}
+}
+
+func slicesContains(s []string, v string) bool {
+	for _, x := range s {
+		if x == v {
+			return true
+		}
+	}
+	return false
+}

management/server/affected_peers_querycount_test.go

@@ -0,0 +1,164 @@
+package server
+
+import (
+	"context"
+	"sync"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	nbdns "github.com/netbirdio/netbird/dns"
+	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
+	resourceTypes "github.com/netbirdio/netbird/management/server/networks/resources/types"
+	routerTypes "github.com/netbirdio/netbird/management/server/networks/routers/types"
+	networkTypes "github.com/netbirdio/netbird/management/server/networks/types"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/route"
+)
+
+// countingStore wraps a real store and counts the per-account collection loads
+// the resolver performs, so a test can assert each is read at most once and that
+// irrelevant collections are skipped entirely.
+type countingStore struct {
+	store.Store
+	mu     sync.Mutex
+	counts map[string]int
+}
+
+func newCountingStore(s store.Store) *countingStore {
+	return &countingStore{Store: s, counts: map[string]int{}}
+}
+
+func (c *countingStore) bump(name string) {
+	c.mu.Lock()
+	c.counts[name]++
+	c.mu.Unlock()
+}
+
+func (c *countingStore) count(name string) int {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	return c.counts[name]
+}
+
+func (c *countingStore) total() int {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	n := 0
+	for _, v := range c.counts {
+		n += v
+	}
+	return n
+}
+
+func (c *countingStore) GetAccountPolicies(ctx context.Context, ls store.LockingStrength, accountID string) ([]*types.Policy, error) {
+	c.bump("policies")
+	return c.Store.GetAccountPolicies(ctx, ls, accountID)
+}
+
+func (c *countingStore) GetAccountRoutes(ctx context.Context, ls store.LockingStrength, accountID string) ([]*route.Route, error) {
+	c.bump("routes")
+	return c.Store.GetAccountRoutes(ctx, ls, accountID)
+}
+
+func (c *countingStore) GetAccountNameServerGroups(ctx context.Context, ls store.LockingStrength, accountID string) ([]*nbdns.NameServerGroup, error) {
+	c.bump("nameservers")
+	return c.Store.GetAccountNameServerGroups(ctx, ls, accountID)
+}
+
+func (c *countingStore) GetAccountDNSSettings(ctx context.Context, ls store.LockingStrength, accountID string) (*types.DNSSettings, error) {
+	c.bump("dnssettings")
+	return c.Store.GetAccountDNSSettings(ctx, ls, accountID)
+}
+
+func (c *countingStore) GetNetworkRoutersByAccountID(ctx context.Context, ls store.LockingStrength, accountID string) ([]*routerTypes.NetworkRouter, error) {
+	c.bump("routers")
+	return c.Store.GetNetworkRoutersByAccountID(ctx, ls, accountID)
+}
+
+func (c *countingStore) GetNetworkResourcesByAccountID(ctx context.Context, ls store.LockingStrength, accountID string) ([]*resourceTypes.NetworkResource, error) {
+	c.bump("resources")
+	return c.Store.GetNetworkResourcesByAccountID(ctx, ls, accountID)
+}
+
+func (c *countingStore) GetAccountServices(ctx context.Context, ls store.LockingStrength, accountID string) ([]*rpservice.Service, error) {
+	c.bump("services")
+	return c.Store.GetAccountServices(ctx, ls, accountID)
+}
+
+// TestAffectedPeers_QueryCount_NoRedundantFullTableLoads asserts the resolver
+// loads each per-account collection at most once per Resolve (memoization) even
+// on a change that drives every bridge, and skips the services table when the
+// account has no embedded proxy peers.
+func TestAffectedPeers_QueryCount_NoRedundantFullTableLoads(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	cs := newCountingStore(s.manager.Store)
+
+	// A group change that exercises policies, routers, resources and the bridge.
+	change := affectedpeers.Change{ChangedGroupIDs: []string{s.sourceGroupID}}
+	snap, err := affectedpeers.Load(ctx, cs, s.accountID, change)
+	require.NoError(t, err)
+	affected := snap.Expand(ctx, s.accountID, change)
+	assert.Contains(t, affected, s.routerPeerID, "bridge must still resolve the routing peer")
+
+	for _, name := range []string{"policies", "routes", "nameservers", "dnssettings", "routers", "resources"} {
+		assert.LessOrEqualf(t, cs.count(name), 1,
+			"%s must be loaded at most once per Resolve, got %d", name, cs.count(name))
+	}
+	assert.Equal(t, 0, cs.count("services"),
+		"services must not be loaded when the account has no embedded proxy peers")
+}
+
+// TestAffectedPeers_QueryCount_NarrowChangeSkipsLoads asserts that a change with
+// no group/peer signal touches no per-account collections beyond what its inputs
+// require.
+func TestAffectedPeers_QueryCount_NarrowChangeSkipsLoads(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	cs := newCountingStore(s.manager.Store)
+
+	// A bare network change drives only the router->source bridge: routers and
+	// resources are needed, but routes/nameservers/dnssettings/services are not.
+	_, err := affectedpeers.Load(ctx, cs, s.accountID, affectedpeers.Change{Networks: []*networkTypes.Network{{ID: s.networkID}}})
+	require.NoError(t, err)
+
+	assert.Equal(t, 0, cs.count("routes"), "routes must not be loaded for a network-only change")
+	assert.Equal(t, 0, cs.count("nameservers"), "nameservers must not be loaded for a network-only change")
+	assert.Equal(t, 0, cs.count("dnssettings"), "dnssettings must not be loaded for a network-only change")
+	assert.Equal(t, 0, cs.count("services"), "services must not be loaded for a network-only change")
+}
+
+// TestAffectedPeers_QueryCount_ExpandReadsNothing is the core invariant of the
+// Load/Expand split: Load (run inside the transaction) does all store reads;
+// Expand (run after commit) must touch the store ZERO times, so it never holds
+// the write lock and never reads post-commit state.
+func TestAffectedPeers_QueryCount_ExpandReadsNothing(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	change := affectedpeers.Change{ChangedGroupIDs: []string{s.sourceGroupID}}
+
+	cs := newCountingStore(s.manager.Store)
+	snap, err := affectedpeers.Load(ctx, cs, s.accountID, change)
+	require.NoError(t, err)
+	require.Greater(t, cs.total(), 0, "Load must read the store")
+
+	// Any store access during Expand would increment the same counter. Expand
+	// operates purely on the snapshot, so the count must not move.
+	readsAfterLoad := cs.total()
+	affected := snap.Expand(ctx, s.accountID, change)
+	assert.Contains(t, affected, s.routerPeerID, "Expand must still produce the affected peers from the snapshot")
+	assert.Equal(t, readsAfterLoad, cs.total(), "Expand must perform zero store reads — it operates purely on the loaded snapshot")
+}

management/server/affected_peers_router_paths_test.go

@@ -0,0 +1,398 @@
+package server
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
+	resourceTypes "github.com/netbirdio/netbird/management/server/networks/resources/types"
+	routerTypes "github.com/netbirdio/netbird/management/server/networks/routers/types"
+	"github.com/netbirdio/netbird/management/server/posture"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+func (s *routerScenario) resolveGroupChangeAffected(ctx context.Context, changedGroupIDs []string) []string {
+	change := affectedpeers.Change{ChangedGroupIDs: changedGroupIDs}
+	snap, err := affectedpeers.Load(ctx, s.manager.Store, s.accountID, change)
+	if err != nil {
+		return nil
+	}
+	return snap.Expand(ctx, s.accountID, change)
+}
+
+func (s *routerScenario) resolvePeerChangeAffected(ctx context.Context, changedPeerIDs []string) []string {
+	change := affectedpeers.Change{ChangedPeerIDs: changedPeerIDs}
+	snap, err := affectedpeers.Load(ctx, s.manager.Store, s.accountID, change)
+	if err != nil {
+		return nil
+	}
+	return snap.Expand(ctx, s.accountID, change)
+}
+
+func TestAffectedPeers_GroupChange_SourceGroupMembership_RefreshesRoutingPeer_DirectRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	affected := s.resolveGroupChangeAffected(ctx, []string{s.sourceGroupID})
+
+	assert.Contains(t, affected, s.sourcePeerID, "source group member must be affected")
+	assert.Contains(t, affected, s.routerPeerID,
+		"changing the source group of a peer->resource policy must refresh the resource's routing peer")
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func TestAffectedPeers_GroupChange_SourceGroupMembership_RefreshesRoutingPeer_RouterPeerGroups(t *testing.T) {
+	s := setupRouterScenario(t, false)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	affected := s.resolveGroupChangeAffected(ctx, []string{s.sourceGroupID})
+
+	assert.Contains(t, affected, s.routerGroupPeerID,
+		"changing the source group must refresh the routing peer defined via router.PeerGroups")
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func TestAffectedPeers_GroupChange_RouterPeerGroupMembership_RefreshesPolicySources(t *testing.T) {
+	s := setupRouterScenario(t, false)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	affected := s.resolveGroupChangeAffected(ctx, []string{s.routerPeerGroupID})
+
+	assert.Contains(t, affected, s.routerGroupPeerID, "the routing peer itself must be affected")
+	assert.Contains(t, affected, s.sourcePeerID,
+		"changing the router's PeerGroups must refresh the source peers of policies serving the resource")
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func TestAffectedPeers_PeerChange_SourcePeer_RefreshesRoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	affected := s.resolvePeerChangeAffected(ctx, []string{s.sourcePeerID})
+
+	assert.Contains(t, affected, s.routerPeerID,
+		"a status change on a source peer must refresh the resource's routing peer that serves it")
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func TestAffectedPeers_PeerChange_SourcePeer_ByDestinationResource_RefreshesRoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID), true)
+	require.NoError(t, err)
+
+	affected := s.resolvePeerChangeAffected(ctx, []string{s.sourcePeerID})
+
+	assert.Contains(t, affected, s.routerPeerID,
+		"DestinationResource-targeted policy must still bridge a source-peer change to the routing peer")
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func TestAffectedPeers_E2E_DeleteGroup_ResolvesAffectedPeers(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	const memberOnlyGroupID = "rs-memberonly-grp"
+	require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{
+		ID: memberOnlyGroupID, Name: "rs-memberonly", Peers: []string{s.sourcePeerID},
+	}))
+
+	affected := s.resolveGroupChangeAffected(ctx, []string{memberOnlyGroupID})
+	assert.Empty(t, affected, "an unlinked group has no network-map impact, so no peer is affected")
+
+	require.NoError(t, s.manager.DeleteGroup(ctx, s.accountID, userID, memberOnlyGroupID))
+}
+
+func TestAffectedPeers_GroupAddResource_RefreshesRoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	const extraResourceGroupID = "rs-resource-grp-extra"
+	require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{
+		ID: extraResourceGroupID, Name: "rs-resource-extra",
+	}))
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, extraResourceGroupID), true)
+	require.NoError(t, err)
+
+	require.NoError(t, s.manager.GroupAddResource(ctx, s.accountID, extraResourceGroupID, types.Resource{
+		ID:   s.resourceID,
+		Type: types.ResourceTypeHost,
+	}))
+
+	affected := s.resolveGroupChangeAffected(ctx, []string{extraResourceGroupID})
+
+	assert.Contains(t, affected, s.routerPeerID,
+		"attaching a resource to a policy destination group must refresh the resource's routing peer")
+	assert.Contains(t, affected, s.sourcePeerID, "policy source peers must refresh")
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func (s *routerScenario) createPostureCheckGatedPolicy(t *testing.T, ctx context.Context, policy *types.Policy) string {
+	t.Helper()
+
+	check, err := s.manager.SavePostureChecks(ctx, s.accountID, userID, &posture.Checks{
+		Name: "rs-min-version",
+		Checks: posture.ChecksDefinition{
+			NBVersionCheck: &posture.NBVersionCheck{MinVersion: "0.30.0"},
+		},
+	}, true)
+	require.NoError(t, err)
+
+	policy.SourcePostureChecks = []string{check.ID}
+	_, err = s.manager.SavePolicy(ctx, s.accountID, userID, policy, true)
+	require.NoError(t, err)
+
+	return check.ID
+}
+
+func TestAffectedPeers_E2E_SavePostureCheck_RefreshesRoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	checkID := s.createPostureCheckGatedPolicy(t, ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	unrelatedCh := s.updateManager.CreateChannel(ctx, s.unrelatedPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+		s.updateManager.CloseChannel(ctx, s.unrelatedPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh, unrelatedCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		peerShouldNotReceiveUpdate(t, unrelatedCh)
+		close(done)
+	}()
+
+	_, err := s.manager.SavePostureChecks(ctx, s.accountID, userID, &posture.Checks{
+		ID:   checkID,
+		Name: "rs-min-version",
+		Checks: posture.ChecksDefinition{
+			NBVersionCheck: &posture.NBVersionCheck{MinVersion: "0.31.0"},
+		},
+	}, false)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: editing a posture check did not refresh source + routing peers")
+	}
+}
+
+func TestAffectedPeers_E2E_UpdateResource_DestinationResourcePolicy_RefreshesSourcePeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID), true)
+	require.NoError(t, err)
+
+	resourcesManager, _, _ := s.managers()
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	unrelatedCh := s.updateManager.CreateChannel(ctx, s.unrelatedPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+		s.updateManager.CloseChannel(ctx, s.unrelatedPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh, unrelatedCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		peerShouldNotReceiveUpdate(t, unrelatedCh)
+		close(done)
+	}()
+
+	_, err = resourcesManager.UpdateResource(ctx, userID, &resourceTypes.NetworkResource{
+		ID:        s.resourceID,
+		AccountID: s.accountID,
+		NetworkID: s.networkID,
+		Name:      "rs-resource-host",
+		Address:   "10.20.30.0/25",
+		GroupIDs:  []string{s.resourceGroupID},
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: updating a DestinationResource-targeted resource did not refresh its policy source peer")
+	}
+}
+
+// A disabled sibling router routes to nobody, so updating a resource on its network
+// must NOT refresh its peer (the enabled router carries the bridge instead).
+func TestAffectedPeers_E2E_UpdateResource_DisabledSiblingRouterNotBridged(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	resourcesManager, routersManager, _ := s.managers()
+
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "rs-key-disabled", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	disabledRouterPeer := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+	_, err = routersManager.CreateRouter(ctx, userID, &routerTypes.NetworkRouter{
+		NetworkID:  s.networkID,
+		AccountID:  s.accountID,
+		Peer:       disabledRouterPeer.ID,
+		Masquerade: true,
+		Metric:     9000,
+		Enabled:    false,
+	})
+	require.NoError(t, err)
+
+	disabledCh := s.updateManager.CreateChannel(ctx, disabledRouterPeer.ID)
+	enabledCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, disabledRouterPeer.ID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+	})
+
+	settleAffectedUpdates(disabledCh, enabledCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, enabledCh)
+		peerShouldNotReceiveUpdate(t, disabledCh)
+		close(done)
+	}()
+
+	_, err = resourcesManager.UpdateResource(ctx, userID, &resourceTypes.NetworkResource{
+		ID:        s.resourceID,
+		AccountID: s.accountID,
+		NetworkID: s.networkID,
+		Name:      "rs-resource-host",
+		Address:   "10.20.30.0/25",
+		GroupIDs:  []string{s.resourceGroupID},
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout")
+	}
+}
+
+func TestAffectedPeers_GroupChange_RouterInOtherNetworkNotAffected(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	second := s.addSecondTopology(t, "groupiso")
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	affected := s.resolveGroupChangeAffected(ctx, []string{s.sourceGroupID})
+
+	assert.Contains(t, affected, s.routerPeerID, "network A's routing peer must be affected")
+	assert.NotContains(t, affected, second.routerPeerID,
+		"a router in an unrelated network must not be affected by a source-group change for another resource")
+}
+
+func TestAffectedPeers_PeerChange_RouterInOtherNetworkNotAffected(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	second := s.addSecondTopology(t, "peeriso")
+	ctx := context.Background()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	affected := s.resolvePeerChangeAffected(ctx, []string{s.sourcePeerID})
+
+	assert.Contains(t, affected, s.routerPeerID, "network A's routing peer must be affected")
+	assert.NotContains(t, affected, second.routerPeerID,
+		"a router in an unrelated network must not be affected by a source-peer change for another resource")
+}
+
+// TestAffectedPeers_E2E_PostureFlip_RefreshesRoutingPeer drives the customer path
+// on the twin store: the source peer's metadata flips a posture verdict on sync,
+// and the routing peer serving the gated resource must be refreshed in both
+// directions. Without the flip detection the deny direction takes the nmap
+// shortcut (the denied peer's map holds no router) and the allow direction
+// depends on which meta field moved, leaving the routers with a stale map.
+func TestAffectedPeers_E2E_PostureFlip_RefreshesRoutingPeer(t *testing.T) {
+	runPostureFlipRefreshesRoutingPeer(t, func(s *routerScenario) *types.Policy {
+		return peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+	})
+}
+
+// TestAffectedPeers_E2E_PostureFlip_DirectSourcePeer_RefreshesRoutingPeer is the same
+// scenario with the source peer named directly in the rule: it must receive its posture
+// checks and have its flips detected exactly like a group member.
+func TestAffectedPeers_E2E_PostureFlip_DirectSourcePeer_RefreshesRoutingPeer(t *testing.T) {
+	runPostureFlipRefreshesRoutingPeer(t, func(s *routerScenario) *types.Policy {
+		return peerToResourcePolicyByPeer(s.sourcePeerID, s.resourceGroupID)
+	})
+}
+
+func runPostureFlipRefreshesRoutingPeer(t *testing.T, policyFor func(s *routerScenario) *types.Policy) {
+	t.Helper()
+
+	manager, updateManager := createManagerWithNetworkMapStore(t)
+	s := buildRouterScenario(t, manager, updateManager, true)
+	ctx := context.Background()
+
+	s.createPostureCheckGatedPolicy(t, ctx, policyFor(s))
+
+	source, err := s.manager.Store.GetPeerByID(ctx, store.LockingStrengthNone, s.accountID, s.sourcePeerID)
+	require.NoError(t, err)
+
+	syncWithVersion := func(version string) {
+		meta := source.Meta
+		meta.WtVersion = version
+		_, _, _, _, err := s.manager.SyncPeer(ctx, types.PeerSync{WireGuardPubKey: source.Key, Meta: meta}, s.accountID)
+		require.NoError(t, err)
+	}
+	syncWithVersion("0.31.0")
+
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	unrelatedCh := s.updateManager.CreateChannel(ctx, s.unrelatedPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+		s.updateManager.CloseChannel(ctx, s.unrelatedPeerID)
+	})
+	settleAffectedUpdates(routerCh, unrelatedCh)
+
+	syncWithVersion("0.29.0")
+	peerShouldReceiveUpdate(t, routerCh)
+	peerShouldNotReceiveUpdate(t, unrelatedCh)
+
+	syncWithVersion("0.31.0")
+	peerShouldReceiveUpdate(t, routerCh)
+	peerShouldNotReceiveUpdate(t, unrelatedCh)
+}

management/server/affected_peers_router_test.go

@@ -0,0 +1,795 @@
+package server
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map"
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
+	"github.com/netbirdio/netbird/management/server/groups"
+	"github.com/netbirdio/netbird/management/server/networks"
+	"github.com/netbirdio/netbird/management/server/networks/resources"
+	resourceTypes "github.com/netbirdio/netbird/management/server/networks/resources/types"
+	"github.com/netbirdio/netbird/management/server/networks/routers"
+	routerTypes "github.com/netbirdio/netbird/management/server/networks/routers/types"
+	networkTypes "github.com/netbirdio/netbird/management/server/networks/types"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// routerScenario captures the topology from the bug report:
+//
+//	network ── router (routing peer) ── resource (in resourceGroup)
+//	independent peer ──(policy: source -> resource)──> resource
+//
+// The routing peer must be refreshed when a policy grants a source peer access
+// to the resource, because the network map connects the source peer to the
+// routing peer at compute time (Account.GetPoliciesForNetworkResource +
+// addNetworksRoutingPeers). The routing peer is NOT a member of the resource
+// group, so static group/peer resolution alone cannot find it.
+type routerScenario struct {
+	manager       *DefaultAccountManager
+	updateManager *update_channel.PeersUpdateManager
+	accountID     string
+	networkID     string
+
+	sourcePeerID  string // independent peer that the policy grants access from
+	sourceGroupID string // group containing the source peer
+
+	routerPeerID      string // peer acting as the routing peer (direct router.Peer)
+	routerGroupPeerID string // peer that is a member of routerPeerGroup
+	routerPeerGroupID string // group used for router.PeerGroups
+
+	resourceID      string // network resource
+	resourceGroupID string // group whose member is the resource (no peers)
+
+	unrelatedPeerID string // peer in no relevant entity
+}
+
+// setupRouterScenario builds the topology above with the default policy removed
+// and channels NOT yet created, so callers control exactly when updates can flow.
+func setupRouterScenario(t *testing.T, directRouterPeer bool) *routerScenario {
+	t.Helper()
+
+	manager, updateManager, err := createManager(t)
+	require.NoError(t, err)
+
+	return buildRouterScenario(t, manager, updateManager, directRouterPeer)
+}
+
+func buildRouterScenario(t *testing.T, manager *DefaultAccountManager, updateManager *update_channel.PeersUpdateManager, directRouterPeer bool) *routerScenario {
+	t.Helper()
+
+	ctx := context.Background()
+
+	account, err := createAccount(manager, "router_scenario", userID, "")
+	require.NoError(t, err)
+	accountID := account.Id
+
+	// Remove the default policy so AddPeer/CreateGroup don't schedule unrelated updates.
+	policies, err := manager.Store.GetAccountPolicies(ctx, store.LockingStrengthNone, accountID)
+	require.NoError(t, err)
+	for _, p := range policies {
+		require.NoError(t, manager.Store.DeletePolicy(ctx, accountID, p.ID))
+	}
+
+	setupKey, err := manager.CreateSetupKey(ctx, accountID, "rs-key", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+
+	sourcePeer := addPeerToAccount(t, manager, accountID, setupKey.Key)
+	routerPeer := addPeerToAccount(t, manager, accountID, setupKey.Key)
+	routerGroupPeer := addPeerToAccount(t, manager, accountID, setupKey.Key)
+	unrelatedPeer := addPeerToAccount(t, manager, accountID, setupKey.Key)
+
+	const (
+		sourceGroupID     = "rs-source-grp"
+		routerPeerGroupID = "rs-router-grp"
+		resourceGroupID   = "rs-resource-grp"
+	)
+
+	for _, g := range []*types.Group{
+		{ID: sourceGroupID, Name: "rs-source", Peers: []string{sourcePeer.ID}},
+		{ID: routerPeerGroupID, Name: "rs-router", Peers: []string{routerGroupPeer.ID}},
+		{ID: resourceGroupID, Name: "rs-resource"}, // intentionally peerless; the resource is its only member
+	} {
+		require.NoError(t, manager.CreateGroup(ctx, accountID, userID, g))
+	}
+
+	groupsManager := groups.NewManager(manager.Store, manager)
+	resourcesManager := resources.NewManager(manager.Store, groupsManager, manager, manager.serviceManager)
+	routersManager := routers.NewManager(manager.Store, manager)
+	networksManager := networks.NewManager(manager.Store, resourcesManager, routersManager, manager)
+
+	network, err := networksManager.CreateNetwork(ctx, userID, &networkTypes.Network{
+		ID:        "rs-network",
+		AccountID: accountID,
+		Name:      "rs-network",
+	})
+	require.NoError(t, err)
+
+	resource, err := resourcesManager.CreateResource(ctx, userID, &resourceTypes.NetworkResource{
+		AccountID: accountID,
+		NetworkID: network.ID,
+		Name:      "rs-resource-host",
+		Address:   "10.20.30.0/24",
+		GroupIDs:  []string{resourceGroupID},
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	router := &routerTypes.NetworkRouter{
+		ID:         "rs-router",
+		NetworkID:  network.ID,
+		AccountID:  accountID,
+		Masquerade: true,
+		Metric:     9999,
+		Enabled:    true,
+	}
+	if directRouterPeer {
+		router.Peer = routerPeer.ID
+	} else {
+		router.PeerGroups = []string{routerPeerGroupID}
+	}
+	_, err = routersManager.CreateRouter(ctx, userID, router)
+	require.NoError(t, err)
+
+	return &routerScenario{
+		manager:           manager,
+		updateManager:     updateManager,
+		accountID:         accountID,
+		networkID:         network.ID,
+		sourcePeerID:      sourcePeer.ID,
+		sourceGroupID:     sourceGroupID,
+		routerPeerID:      routerPeer.ID,
+		routerGroupPeerID: routerGroupPeer.ID,
+		routerPeerGroupID: routerPeerGroupID,
+		resourceID:        resource.ID,
+		resourceGroupID:   resourceGroupID,
+		unrelatedPeerID:   unrelatedPeer.ID,
+	}
+}
+
+// peerToResourcePolicy builds a policy granting the source group access to the
+// resource, referencing the resource by its group in the rule destination.
+func peerToResourcePolicyByGroup(sourceGroupID, resourceGroupID string) *types.Policy {
+	return &types.Policy{
+		Enabled: true,
+		Name:    "peer-to-resource-by-group",
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:      true,
+				Sources:      []string{sourceGroupID},
+				Destinations: []string{resourceGroupID},
+				Action:       types.PolicyTrafficActionAccept,
+			},
+		},
+	}
+}
+
+// peerToResourcePolicyByPeer builds a policy naming the source peer directly via
+// SourceResource rather than through a group.
+func peerToResourcePolicyByPeer(sourcePeerID, resourceGroupID string) *types.Policy {
+	return &types.Policy{
+		Enabled: true,
+		Name:    "peer-to-resource-by-peer",
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:        true,
+				SourceResource: types.Resource{ID: sourcePeerID, Type: types.ResourceTypePeer},
+				Destinations:   []string{resourceGroupID},
+				Action:         types.PolicyTrafficActionAccept,
+			},
+		},
+	}
+}
+
+// peerToResourcePolicyByResource builds a policy referencing the resource
+// directly via DestinationResource rather than its group.
+func peerToResourcePolicyByResource(sourceGroupID, resourceID string) *types.Policy {
+	return &types.Policy{
+		Enabled: true,
+		Name:    "peer-to-resource-by-resource",
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:             true,
+				Sources:             []string{sourceGroupID},
+				DestinationResource: types.Resource{ID: resourceID, Type: types.ResourceTypeHost},
+				Action:              types.PolicyTrafficActionAccept,
+			},
+		},
+	}
+}
+
+// resolvePolicyAffected mirrors SavePolicy's resolution: resolve the affected
+// peers for the given policy.
+func (s *routerScenario) resolvePolicyAffected(ctx context.Context, policy *types.Policy) []string {
+	change := affectedpeers.Change{Policies: []*types.Policy{policy}}
+	snap, err := affectedpeers.Load(ctx, s.manager.Store, s.accountID, change)
+	if err != nil {
+		return nil
+	}
+	return snap.Expand(ctx, s.accountID, change)
+}
+
+func TestAffectedPeers_SourcePeer_DirectRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	assert.Contains(t, affected, s.sourcePeerID, "source peer must be affected")
+}
+
+func TestAffectedPeers_RoutingPeer_DirectRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	// BUG: the direct routing peer serves the resource's subnet to the source
+	// peer, so it must be refreshed when the policy is created. The policy path
+	// only resolves the literal rule groups (source group + resource group);
+	// the resource group has no peer members and the router peer is reachable
+	// only through the network, so it is dropped.
+	assert.Contains(t, affected, s.routerPeerID,
+		"routing peer (router.Peer) serving the resource must be affected by a policy granting access to it")
+}
+
+func TestAffectedPeers_RoutingPeer_RouterPeerGroups(t *testing.T) {
+	s := setupRouterScenario(t, false)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	// Router defined via PeerGroups instead of a direct peer.
+	assert.Contains(t, affected, s.routerGroupPeerID,
+		"routing peer (router.PeerGroups member) serving the resource must be affected")
+}
+
+func TestAffectedPeers_DestResource_RoutingPeer_DirectRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID)
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	// When the resource is referenced via DestinationResource, RuleGroups()
+	// returns only the source group and the resource ID is not a peer, so
+	// collectPolicyAffectedGroupsAndPeers yields nothing for the destination at
+	// all. The routing peer is dropped here too.
+	assert.Contains(t, affected, s.routerPeerID,
+		"routing peer must be affected when the resource is referenced via DestinationResource")
+}
+
+func TestAffectedPeers_DestResource_RoutingPeer_RouterPeerGroups(t *testing.T) {
+	s := setupRouterScenario(t, false)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID)
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	assert.Contains(t, affected, s.routerGroupPeerID,
+		"routing peer (PeerGroups) must be affected when the resource is referenced via DestinationResource")
+}
+
+func TestAffectedPeers_SourceResourcePeer_RoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	// Source expressed as a direct peer (SourceResource), destination as resource group.
+	policy := &types.Policy{
+		Enabled: true,
+		Name:    "sourceResource-peer-to-resource",
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:        true,
+				SourceResource: types.Resource{ID: s.sourcePeerID, Type: types.ResourceTypePeer},
+				Destinations:   []string{s.resourceGroupID},
+				Action:         types.PolicyTrafficActionAccept,
+			},
+		},
+	}
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	// The direct source peer IS picked up (collectPolicyAffectedGroupsAndPeers
+	// handles SourceResource peers), but the routing peer is still missing.
+	assert.Contains(t, affected, s.sourcePeerID, "direct source peer must be affected")
+	assert.Contains(t, affected, s.routerPeerID, "routing peer must be affected")
+}
+
+func TestAffectedPeers_PolicyToResource_UnrelatedPeerNotAffected(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	// Guard against an over-broad fix: a peer in no relevant entity must never
+	// be pulled in.
+	assert.NotContains(t, affected, s.unrelatedPeerID, "unrelated peer must not be affected")
+}
+
+func TestAffectedPeers_ResourceSideBridgesToRoutingPeer_DirectRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	// A pre-existing policy grants the source group access to the resource.
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	// Drive an update through the resource manager and assert the routing peer
+	// is among the affected set by observing the channel. This path walks
+	// policies whose destinations reference the resource's groups, folds in the
+	// source groups, and loads the network's routers, so it reaches both the
+	// source peer and the routing peer.
+	groupsManager := groups.NewManager(s.manager.Store, s.manager)
+	rm := resources.NewManager(s.manager.Store, groupsManager, s.manager, s.manager.serviceManager)
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+	})
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		close(done)
+	}()
+
+	_, err = rm.UpdateResource(ctx, userID, &resourceTypes.NetworkResource{
+		ID:        s.resourceID,
+		AccountID: s.accountID,
+		NetworkID: s.networkID,
+		Name:      "rs-resource-host",
+		Address:   "10.20.30.0/24",
+		GroupIDs:  []string{s.resourceGroupID},
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: resource update did not refresh source peer + routing peer")
+	}
+}
+
+// settleAffectedUpdates waits for in-flight async updates to arrive, then drains
+// every given channel so subsequent assertions start from a clean slate.
+//
+// Setup (CreateNetwork/CreateResource/CreateRouter) fires async UpdateAffectedPeers
+// goroutines; draining first means the assertion only observes updates from the
+// action under test, not setup stragglers.
+func settleAffectedUpdates(chans ...<-chan *network_map.UpdateMessage) {
+	time.Sleep(300 * time.Millisecond)
+	for _, ch := range chans {
+		drainPeerUpdates(ch)
+	}
+}
+
+func TestAffectedPeers_E2E_CreatePolicy_RoutingPeer_DirectRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	unrelatedCh := s.updateManager.CreateChannel(ctx, s.unrelatedPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+		s.updateManager.CloseChannel(ctx, s.unrelatedPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh, unrelatedCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		peerShouldNotReceiveUpdate(t, unrelatedCh)
+		close(done)
+	}()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: creating peer->resource policy did not refresh the routing peer")
+	}
+}
+
+func TestAffectedPeers_E2E_CreatePolicy_RoutingPeer_RouterPeerGroups(t *testing.T) {
+	s := setupRouterScenario(t, false)
+	ctx := context.Background()
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerGroupPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerGroupPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		close(done)
+	}()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: routing peer (PeerGroups) not refreshed on policy create")
+	}
+}
+
+func TestAffectedPeers_E2E_DestResource_RoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		close(done)
+	}()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID), true)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: routing peer not refreshed when policy targets DestinationResource")
+	}
+}
+
+func TestAffectedPeers_E2E_DeletePolicy_RoutingPeer(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	policy, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		close(done)
+	}()
+
+	require.NoError(t, s.manager.DeletePolicy(ctx, s.accountID, policy.ID, userID))
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: deleting peer->resource policy did not refresh the routing peer")
+	}
+}
+
+func (s *routerScenario) managers() (resources.Manager, routers.Manager, networks.Manager) {
+	groupsManager := groups.NewManager(s.manager.Store, s.manager)
+	resourcesManager := resources.NewManager(s.manager.Store, groupsManager, s.manager, s.manager.serviceManager)
+	routersManager := routers.NewManager(s.manager.Store, s.manager)
+	networksManager := networks.NewManager(s.manager.Store, resourcesManager, routersManager, s.manager)
+	return resourcesManager, routersManager, networksManager
+}
+
+type secondTopology struct {
+	networkID       string
+	resourceID      string
+	resourceGroupID string
+	routerPeerID    string
+}
+
+func (s *routerScenario) addSecondTopology(t *testing.T, suffix string) secondTopology {
+	t.Helper()
+	ctx := context.Background()
+	resourcesManager, routersManager, networksManager := s.managers()
+
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "rs-key-"+suffix, types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	routerPeer := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+
+	resourceGroupID := "rs-resource-grp-" + suffix
+	require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{
+		ID: resourceGroupID, Name: "rs-resource-" + suffix,
+	}))
+
+	network, err := networksManager.CreateNetwork(ctx, userID, &networkTypes.Network{
+		ID:        "rs-network-" + suffix,
+		AccountID: s.accountID,
+		Name:      "rs-network-" + suffix,
+	})
+	require.NoError(t, err)
+
+	resource, err := resourcesManager.CreateResource(ctx, userID, &resourceTypes.NetworkResource{
+		AccountID: s.accountID,
+		NetworkID: network.ID,
+		Name:      "rs-resource-host-" + suffix,
+		Address:   "10.40.50.0/24",
+		GroupIDs:  []string{resourceGroupID},
+		Enabled:   true,
+	})
+	require.NoError(t, err)
+
+	_, err = routersManager.CreateRouter(ctx, userID, &routerTypes.NetworkRouter{
+		NetworkID:  network.ID,
+		AccountID:  s.accountID,
+		Peer:       routerPeer.ID,
+		Masquerade: true,
+		Metric:     9999,
+		Enabled:    true,
+	})
+	require.NoError(t, err)
+
+	return secondTopology{
+		networkID:       network.ID,
+		resourceID:      resource.ID,
+		resourceGroupID: resourceGroupID,
+		routerPeerID:    routerPeer.ID,
+	}
+}
+
+func TestAffectedPeers_E2E_UpdatePolicy_BothRoutingPeers(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	second := s.addSecondTopology(t, "b")
+	ctx := context.Background()
+
+	policy, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerACh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	routerBCh := s.updateManager.CreateChannel(ctx, second.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+		s.updateManager.CloseChannel(ctx, second.routerPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerACh, routerBCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerACh)
+		peerShouldReceiveUpdate(t, routerBCh)
+		close(done)
+	}()
+
+	policy.Rules[0].Destinations = []string{second.resourceGroupID}
+	_, err = s.manager.SavePolicy(ctx, s.accountID, userID, policy, false)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: re-pointing the policy destination did not refresh both routing peers")
+	}
+}
+
+func TestAffectedPeers_E2E_UpdatePolicy_AddSource(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	const secondSourceGroupID = "rs-source-grp-2"
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "rs-key-2", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	secondSourcePeer := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+	require.NoError(t, s.manager.CreateGroup(ctx, s.accountID, userID, &types.Group{
+		ID: secondSourceGroupID, Name: "rs-source-2", Peers: []string{secondSourcePeer.ID},
+	}))
+
+	policy, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
+	require.NoError(t, err)
+
+	newSrcCh := s.updateManager.CreateChannel(ctx, secondSourcePeer.ID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, secondSourcePeer.ID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+	})
+
+	settleAffectedUpdates(newSrcCh, routerCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, newSrcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		close(done)
+	}()
+
+	policy.Rules[0].Sources = []string{s.sourceGroupID, secondSourceGroupID}
+	_, err = s.manager.SavePolicy(ctx, s.accountID, userID, policy, false)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: adding a source group did not refresh the new source peer + routing peer")
+	}
+}
+
+func TestAffectedPeers_E2E_DestResource_RouterPeerGroups(t *testing.T) {
+	s := setupRouterScenario(t, false)
+	ctx := context.Background()
+
+	srcCh := s.updateManager.CreateChannel(ctx, s.sourcePeerID)
+	routerCh := s.updateManager.CreateChannel(ctx, s.routerGroupPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, s.sourcePeerID)
+		s.updateManager.CloseChannel(ctx, s.routerGroupPeerID)
+	})
+
+	settleAffectedUpdates(srcCh, routerCh)
+
+	done := make(chan struct{})
+	go func() {
+		peerShouldReceiveUpdate(t, srcCh)
+		peerShouldReceiveUpdate(t, routerCh)
+		close(done)
+	}()
+
+	_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByResource(s.sourceGroupID, s.resourceID), true)
+	require.NoError(t, err)
+
+	select {
+	case <-done:
+	case <-time.After(peerUpdateTimeout):
+		t.Error("timeout: DestinationResource policy with PeerGroups router did not refresh the routing peer")
+	}
+}
+
+func TestAffectedPeers_AllRoutingPeers_Network(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	_, routersManager, _ := s.managers()
+	setupKey, err := s.manager.CreateSetupKey(ctx, s.accountID, "rs-key-r2", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+	secondRouterPeer := addPeerToAccount(t, s.manager, s.accountID, setupKey.Key)
+	_, err = routersManager.CreateRouter(ctx, userID, &routerTypes.NetworkRouter{
+		NetworkID:  s.networkID,
+		AccountID:  s.accountID,
+		Peer:       secondRouterPeer.ID,
+		Masquerade: true,
+		Metric:     9998,
+		Enabled:    true,
+	})
+	require.NoError(t, err)
+
+	affected := s.resolvePolicyAffected(ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
+
+	assert.Contains(t, affected, s.routerPeerID, "first routing peer must be affected")
+	assert.Contains(t, affected, secondRouterPeer.ID, "second routing peer on the same network must also be affected")
+}
+
+// A disabled router in the snapshot routes to nobody, so it is skipped when the
+// walk scans existing account data: a policy edit still folds the literal source
+// group, but not the disabled router's peer.
+func TestAffectedPeers_DisabledRouter(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	routers, err := s.manager.Store.GetNetworkRoutersByNetID(ctx, store.LockingStrengthNone, s.accountID, s.networkID)
+	require.NoError(t, err)
+	require.Len(t, routers, 1)
+	routers[0].Enabled = false
+	require.NoError(t, s.manager.Store.UpdateNetworkRouter(ctx, routers[0]))
+
+	affected := s.resolvePolicyAffected(ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
+
+	assert.Contains(t, affected, s.sourcePeerID, "source peer (literal policy source group) must be affected")
+	assert.NotContains(t, affected, s.routerPeerID,
+		"a disabled router routes to nobody, so its peer must not be folded from snapshot data")
+}
+
+// A disabled resource in the snapshot is skipped: the policy edit still folds the
+// literal source group, but the resource no longer bridges to its network's router.
+func TestAffectedPeers_DisabledResource(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	res, err := s.manager.Store.GetNetworkResourceByID(ctx, store.LockingStrengthNone, s.accountID, s.resourceID)
+	require.NoError(t, err)
+	res.Enabled = false
+	require.NoError(t, s.manager.Store.SaveNetworkResource(ctx, res))
+
+	affected := s.resolvePolicyAffected(ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
+
+	assert.Contains(t, affected, s.sourcePeerID, "source peer (literal policy source group) must be affected")
+	assert.NotContains(t, affected, s.routerPeerID,
+		"a disabled resource routes to nobody, so its network's router must not be folded from snapshot data")
+}
+
+func TestAffectedPeers_DisabledRule(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	ctx := context.Background()
+
+	policy := peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID)
+	policy.Rules[0].Enabled = false
+
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	assert.Contains(t, affected, s.routerPeerID,
+		"disabled rule must still resolve the routing peer: Enabled must not gate affected-peers")
+}
+
+func TestAffectedPeers_MultiRule(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	second := s.addSecondTopology(t, "c")
+	ctx := context.Background()
+
+	policy := &types.Policy{
+		Enabled: true,
+		Name:    "multi-rule-two-resources",
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:      true,
+				Sources:      []string{s.sourceGroupID},
+				Destinations: []string{s.resourceGroupID},
+				Action:       types.PolicyTrafficActionAccept,
+			},
+			{
+				Enabled:      true,
+				Sources:      []string{s.sourceGroupID},
+				Destinations: []string{second.resourceGroupID},
+				Action:       types.PolicyTrafficActionAccept,
+			},
+		},
+	}
+
+	affected := s.resolvePolicyAffected(ctx, policy)
+
+	assert.Contains(t, affected, s.routerPeerID, "routing peer for resource A must be affected")
+	assert.Contains(t, affected, second.routerPeerID, "routing peer for resource B must be affected")
+}
+
+func TestAffectedPeers_RouterOtherNetwork(t *testing.T) {
+	s := setupRouterScenario(t, true)
+	second := s.addSecondTopology(t, "d")
+	ctx := context.Background()
+
+	affected := s.resolvePolicyAffected(ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
+
+	assert.Contains(t, affected, s.routerPeerID, "network A's routing peer must be affected")
+	assert.NotContains(t, affected, second.routerPeerID,
+		"a router in an unrelated network must not be affected by a policy that does not target its resource")
+}

management/server/affected_peers_user_test.go

@@ -0,0 +1,170 @@
+package server
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+
+	"github.com/netbirdio/netbird/management/server/activity"
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// A user update refreshes only the peers its auto-group change reaches, and a user
+// update that changes no group membership refreshes nobody.
+func TestAffectedPeers_SaveUser_OnlyAffectedPeersUpdated(t *testing.T) {
+	manager, updateManager, account, _, peer2, peer3 := setupNetworkMapTest(t)
+	ctx := context.Background()
+	accountID := account.Id
+
+	const targetUserID = "target-user"
+	require.NoError(t, manager.Store.SaveUser(ctx, &types.User{
+		Id: targetUserID, AccountID: accountID, Role: types.UserRoleUser,
+	}))
+
+	key, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err)
+	targetPeer, _, _, _, err := manager.AddPeer(ctx, accountID, "", targetUserID, &nbpeer.Peer{
+		Key:  key.PublicKey().String(),
+		Meta: nbpeer.PeerSystemMeta{Hostname: "target-peer"},
+	}, false)
+	require.NoError(t, err)
+
+	policies, err := manager.Store.GetAccountPolicies(ctx, store.LockingStrengthNone, accountID)
+	require.NoError(t, err)
+	for _, p := range policies {
+		require.NoError(t, manager.Store.DeletePolicy(ctx, accountID, p.ID))
+	}
+
+	account, err = manager.Store.GetAccount(ctx, accountID)
+	require.NoError(t, err)
+	account.Settings.GroupsPropagationEnabled = true
+	require.NoError(t, manager.Store.SaveAccount(ctx, account))
+
+	require.NoError(t, manager.CreateGroup(ctx, accountID, userID, &types.Group{ID: "ug-linked", Name: "ug-linked"}))
+	require.NoError(t, manager.CreateGroup(ctx, accountID, userID, &types.Group{ID: "ug-dest", Name: "ug-dest", Peers: []string{peer2.ID}}))
+
+	_, err = manager.SavePolicy(ctx, accountID, userID, &types.Policy{
+		Enabled: true,
+		Rules: []*types.PolicyRule{
+			{
+				Enabled:       true,
+				Sources:       []string{"ug-linked"},
+				Destinations:  []string{"ug-dest"},
+				Bidirectional: true,
+				Action:        types.PolicyTrafficActionAccept,
+			},
+		},
+	}, true)
+	require.NoError(t, err)
+
+	updTarget := updateManager.CreateChannel(ctx, targetPeer.ID)
+	upd2 := updateManager.CreateChannel(ctx, peer2.ID)
+	upd3 := updateManager.CreateChannel(ctx, peer3.ID)
+	t.Cleanup(func() {
+		updateManager.CloseChannel(ctx, targetPeer.ID)
+		updateManager.CloseChannel(ctx, peer2.ID)
+		updateManager.CloseChannel(ctx, peer3.ID)
+	})
+
+	t.Run("auto group change updates only linked peers", func(t *testing.T) {
+		drainPeerUpdates(updTarget)
+		drainPeerUpdates(upd2)
+		drainPeerUpdates(upd3)
+
+		_, err := manager.SaveUser(ctx, accountID, activity.SystemInitiator, &types.User{
+			Id: targetUserID, AccountID: accountID, Role: types.UserRoleUser,
+			AutoGroups: []string{"ug-linked"},
+		})
+		require.NoError(t, err)
+
+		peerShouldReceiveUpdate(t, updTarget)
+		peerShouldReceiveUpdate(t, upd2)
+		peerShouldNotReceiveUpdate(t, upd3)
+	})
+
+	t.Run("update without group changes refreshes nobody", func(t *testing.T) {
+		drainPeerUpdates(updTarget)
+		drainPeerUpdates(upd2)
+		drainPeerUpdates(upd3)
+
+		_, err := manager.SaveUser(ctx, accountID, activity.SystemInitiator, &types.User{
+			Id: targetUserID, AccountID: accountID, Role: types.UserRoleUser,
+			AutoGroups: []string{"ug-linked"}, Name: "renamed",
+		})
+		require.NoError(t, err)
+
+		peerShouldNotReceiveUpdate(t, updTarget)
+		peerShouldNotReceiveUpdate(t, upd2)
+		peerShouldNotReceiveUpdate(t, upd3)
+
+		user, err := manager.Store.GetUserByUserID(ctx, store.LockingStrengthNone, targetUserID)
+		require.NoError(t, err)
+		assert.Equal(t, "renamed", user.Name)
+	})
+
+	t.Run("auto group change reassigning IPv6 refreshes the changed peers and their observers", func(t *testing.T) {
+		account, err := manager.Store.GetAccount(ctx, accountID)
+		require.NoError(t, err)
+		account.Settings.IPv6EnabledGroups = []string{"ug-v6"}
+		require.NoError(t, manager.Store.SaveAccount(ctx, account))
+		require.NoError(t, manager.CreateGroup(ctx, accountID, userID, &types.Group{ID: "ug-v6", Name: "ug-v6"}))
+
+		drainPeerUpdates(updTarget)
+		drainPeerUpdates(upd2)
+		drainPeerUpdates(upd3)
+
+		_, err = manager.SaveUser(ctx, accountID, activity.SystemInitiator, &types.User{
+			Id: targetUserID, AccountID: accountID, Role: types.UserRoleUser,
+			AutoGroups: []string{"ug-linked", "ug-v6"}, Name: "renamed",
+		})
+		require.NoError(t, err)
+
+		// The reassigned peer refreshes with everyone it can reach: peer2 via the
+		// policy, but not peer3, which shares no group or policy with it.
+		peerShouldReceiveUpdate(t, updTarget)
+		peerShouldReceiveUpdate(t, upd2)
+		peerShouldNotReceiveUpdate(t, upd3)
+	})
+
+	t.Run("unblocking a user refreshes only the SSH rule destinations", func(t *testing.T) {
+		// An SSH rule that authorizes no group of its own ships the account's
+		// allowed-user set to its destinations, so those are the peers an unblock
+		// reaches — not the whole account.
+		_, err := manager.SavePolicy(ctx, accountID, userID, &types.Policy{
+			Enabled: true,
+			Rules: []*types.PolicyRule{{
+				Enabled:      true,
+				Sources:      []string{"ug-linked"},
+				Destinations: []string{"ug-dest"},
+				Protocol:     types.PolicyRuleProtocolNetbirdSSH,
+				Action:       types.PolicyTrafficActionAccept,
+			}},
+		}, true)
+		require.NoError(t, err)
+
+		blocked, err := manager.Store.GetUserByUserID(ctx, store.LockingStrengthNone, targetUserID)
+		require.NoError(t, err)
+		blocked.Blocked = true
+		require.NoError(t, manager.Store.SaveUser(ctx, blocked))
+
+		drainPeerUpdates(updTarget)
+		drainPeerUpdates(upd2)
+		drainPeerUpdates(upd3)
+
+		// Same auto-groups as the previous subtest left them, so no group change and
+		// no IPv6 reconciliation interferes: the unblock alone drives the refresh.
+		_, err = manager.SaveUser(ctx, accountID, activity.SystemInitiator, &types.User{
+			Id: targetUserID, AccountID: accountID, Role: types.UserRoleUser,
+			AutoGroups: []string{"ug-linked", "ug-v6"}, Name: "renamed",
+		})
+		require.NoError(t, err)
+
+		peerShouldReceiveUpdate(t, upd2)
+		peerShouldNotReceiveUpdate(t, upd3)
+	})
+}

management/server/affectedpeers/proxy_synth_test.go

@@ -0,0 +1,95 @@
+package affectedpeers
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
+	"github.com/netbirdio/netbird/management/server/store"
+)
+
+// fakeProxyStore implements only the two store methods loadProxyServices calls;
+// the embedded nil store.Store panics if anything else is invoked, which keeps
+// the test honest about the surface under test.
+type fakeProxyStore struct {
+	store.Store
+	proxyByCluster map[string][]string
+	persisted      []*rpservice.Service
+}
+
+func (f *fakeProxyStore) GetEmbeddedProxyPeerIDsByCluster(_ context.Context, _ string) (map[string][]string, error) {
+	return f.proxyByCluster, nil
+}
+
+func (f *fakeProxyStore) GetAccountServices(_ context.Context, _ store.LockingStrength, _ string) ([]*rpservice.Service, error) {
+	return f.persisted, nil
+}
+
+func serviceIDs(svcs []*rpservice.Service) []string {
+	ids := make([]string, 0, len(svcs))
+	for _, s := range svcs {
+		ids = append(ids, s.ID)
+	}
+	return ids
+}
+
+// loadProxyServices must merge the synthesised agent-network services (which are
+// never persisted) with the persisted ones, so the proxy-affected expansion can
+// see agent-network AccessGroups. Without this the embedded proxy peer is never
+// flagged on a client group change and only a full resync (restart) recovers.
+func TestLoadProxyServices_MergesSynthesizedAgentNetworkServices(t *testing.T) {
+	prev := agentNetworkSynthesizer
+	t.Cleanup(func() { agentNetworkSynthesizer = prev })
+	SetAgentNetworkSynthesizer(func(_ context.Context, _ store.Store, _ string) ([]*rpservice.Service, error) {
+		return []*rpservice.Service{
+			{ID: "agent-net-svc-acc", ProxyCluster: "proxy.netbird.local", Private: true, AccessGroups: []string{"gB"}},
+		}, nil
+	})
+
+	s := &fakeProxyStore{
+		proxyByCluster: map[string][]string{"proxy.netbird.local": {"proxy-peer-1"}},
+		persisted:      []*rpservice.Service{{ID: "persisted-rp-svc", ProxyCluster: "proxy.netbird.local"}},
+	}
+	snap := &Snapshot{}
+	require.NoError(t, snap.loadProxyServices(context.Background(), s, "acc"))
+
+	ids := serviceIDs(snap.services)
+	assert.Contains(t, ids, "persisted-rp-svc", "persisted services must be kept")
+	assert.Contains(t, ids, "agent-net-svc-acc", "synthesised agent-network service must be merged in")
+}
+
+// With no synthesiser registered, loadProxyServices falls back to persisted
+// services only (no panic, no behaviour change for non-agent-network builds).
+func TestLoadProxyServices_NoSynthesizerRegistered(t *testing.T) {
+	prev := agentNetworkSynthesizer
+	t.Cleanup(func() { agentNetworkSynthesizer = prev })
+	agentNetworkSynthesizer = nil
+
+	s := &fakeProxyStore{
+		proxyByCluster: map[string][]string{"c": {"proxy-1"}},
+		persisted:      []*rpservice.Service{{ID: "persisted"}},
+	}
+	snap := &Snapshot{}
+	require.NoError(t, snap.loadProxyServices(context.Background(), s, "acc"))
+	assert.Equal(t, []string{"persisted"}, serviceIDs(snap.services))
+}
+
+// No embedded proxy peers → skip entirely (don't even call the synthesiser).
+func TestLoadProxyServices_NoEmbeddedProxyPeersSkips(t *testing.T) {
+	prev := agentNetworkSynthesizer
+	t.Cleanup(func() { agentNetworkSynthesizer = prev })
+	called := false
+	SetAgentNetworkSynthesizer(func(_ context.Context, _ store.Store, _ string) ([]*rpservice.Service, error) {
+		called = true
+		return nil, nil
+	})
+
+	s := &fakeProxyStore{proxyByCluster: map[string][]string{}}
+	snap := &Snapshot{}
+	require.NoError(t, snap.loadProxyServices(context.Background(), s, "acc"))
+	assert.False(t, called, "synthesiser must not run for accounts without embedded proxy peers")
+	assert.Empty(t, snap.services)
+}

management/server/affectedpeers/resolver_test.go

@@ -0,0 +1,114 @@
+package affectedpeers
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+
+	resourceTypes "github.com/netbirdio/netbird/management/server/networks/resources/types"
+	networkTypes "github.com/netbirdio/netbird/management/server/networks/types"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// policyGroupsAndPeers mirrors the both-sides extraction (RuleGroups + direct peers)
+// the resolver folds in for a changed policy, for asserting the pure logic.
+func policyGroupsAndPeers(policies ...*types.Policy) (groups []string, peers []string) {
+	peerSet := map[string]struct{}{}
+	for _, p := range policies {
+		if p == nil {
+			continue
+		}
+		groups = append(groups, p.RuleGroups()...)
+		for _, rule := range p.Rules {
+			if rule.SourceResource.Type == types.ResourceTypePeer && rule.SourceResource.ID != "" {
+				peerSet[rule.SourceResource.ID] = struct{}{}
+			}
+			if rule.DestinationResource.Type == types.ResourceTypePeer && rule.DestinationResource.ID != "" {
+				peerSet[rule.DestinationResource.ID] = struct{}{}
+			}
+		}
+	}
+	for id := range peerSet {
+		peers = append(peers, id)
+	}
+	return groups, peers
+}
+
+func TestPolicyGroupsAndPeers_Basic(t *testing.T) {
+	policy := &types.Policy{Rules: []*types.PolicyRule{{Sources: []string{"g1", "g2"}, Destinations: []string{"g3"}}}}
+	groups, peers := policyGroupsAndPeers(policy)
+	assert.ElementsMatch(t, []string{"g1", "g2", "g3"}, groups)
+	assert.Empty(t, peers)
+}
+
+func TestPolicyGroupsAndPeers_WithPeerResources(t *testing.T) {
+	policy := &types.Policy{Rules: []*types.PolicyRule{{
+		Sources:             []string{"g1"},
+		SourceResource:      types.Resource{ID: "p1", Type: types.ResourceTypePeer},
+		Destinations:        []string{"g2"},
+		DestinationResource: types.Resource{ID: "p2", Type: types.ResourceTypePeer},
+	}}}
+	groups, peers := policyGroupsAndPeers(policy)
+	assert.ElementsMatch(t, []string{"g1", "g2"}, groups)
+	assert.ElementsMatch(t, []string{"p1", "p2"}, peers)
+}
+
+func TestPolicyGroupsAndPeers_NilPolicy(t *testing.T) {
+	groups, peers := policyGroupsAndPeers(nil)
+	assert.Nil(t, groups)
+	assert.Nil(t, peers)
+}
+
+func TestPolicyGroupsAndPeers_MultiplePolicies(t *testing.T) {
+	old := &types.Policy{Rules: []*types.PolicyRule{{Sources: []string{"g1"}, Destinations: []string{"g2"}}}}
+	updated := &types.Policy{Rules: []*types.PolicyRule{{Sources: []string{"g3"}, Destinations: []string{"g4"}}}}
+	groups, _ := policyGroupsAndPeers(updated, old)
+	assert.ElementsMatch(t, []string{"g1", "g2", "g3", "g4"}, groups)
+}
+
+func TestPolicyGroupsAndPeers_NonPeerResource(t *testing.T) {
+	policy := &types.Policy{Rules: []*types.PolicyRule{{
+		Sources:        []string{"g1"},
+		SourceResource: types.Resource{ID: "domain-1", Type: types.ResourceTypeDomain},
+		Destinations:   []string{"g2"},
+	}}}
+	groups, peers := policyGroupsAndPeers(policy)
+	assert.ElementsMatch(t, []string{"g1", "g2"}, groups)
+	assert.Empty(t, peers, "domain resource type should not produce direct peer IDs")
+}
+
+func TestChangeIsEmpty(t *testing.T) {
+	assert.True(t, Change{}.isEmpty())
+	assert.False(t, Change{ChangedGroupIDs: []string{"g"}}.isEmpty())
+	assert.False(t, Change{ChangedPeerIDs: []string{"p"}}.isEmpty())
+	assert.False(t, Change{Policies: []*types.Policy{{}}}.isEmpty())
+	assert.False(t, Change{Resources: []*resourceTypes.NetworkResource{{ID: "r"}}}.isEmpty())
+	assert.False(t, Change{Networks: []*networkTypes.Network{{ID: "n"}}}.isEmpty())
+	assert.False(t, Change{PostureCheckIDs: []string{"pc"}}.isEmpty())
+	assert.False(t, Change{UserGroupIDs: []string{"g"}}.isEmpty())
+	assert.False(t, Change{AllowedUsersChanged: true}.isEmpty())
+}
+
+func TestPolicyReferencesPostureChecks(t *testing.T) {
+	policy := &types.Policy{SourcePostureChecks: []string{"pc1", "pc2"}}
+
+	assert.True(t, policyReferencesPostureChecks(policy, map[string]struct{}{"pc1": {}}))
+	assert.False(t, policyReferencesPostureChecks(policy, map[string]struct{}{"pc3": {}}))
+}
+
+func TestCollectPolicySources(t *testing.T) {
+	policy := &types.Policy{Rules: []*types.PolicyRule{{
+		Enabled:        true,
+		Sources:        []string{"g1"},
+		SourceResource: types.Resource{Type: types.ResourceTypePeer, ID: "p1"},
+		Destinations:   []string{"g2"},
+	}}}
+
+	groupSet := map[string]struct{}{}
+	peerSet := map[string]struct{}{}
+	collectPolicySources(policy, groupSet, peerSet)
+
+	assert.Contains(t, groupSet, "g1")
+	assert.NotContains(t, groupSet, "g2", "destination groups must not be collected as sources")
+	assert.Contains(t, peerSet, "p1")
+}

management/server/agentnetwork_budgetrule_realstack_test.go

@@ -0,0 +1,170 @@
+package server
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
+	agenttypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	"github.com/netbirdio/netbird/management/internals/modules/permissions"
+	"github.com/netbirdio/netbird/management/server/store"
+)
+
+// TestAgentNetwork_BudgetRuleCRUD_RealManager is the GC-1 no-mock guard for the
+// account budget-rule manager surface: real DefaultAccountManager, real store,
+// real permissions. It exercises create/get/list/update/delete through the
+// permission-gated manager (not the store directly) and asserts the reused
+// PolicyLimits cap shape and targets survive each step.
+func TestAgentNetwork_BudgetRuleCRUD_RealManager(t *testing.T) {
+	am, _, err := createManager(t)
+	require.NoError(t, err, "createManager must succeed")
+	ctx := context.Background()
+
+	const (
+		accountID   = "agent-net-budget-acct"
+		adminUserID = "agent-net-budget-admin"
+	)
+	account := newAccountWithId(ctx, accountID, adminUserID, "agent-net.test", "", "", false)
+	require.NoError(t, am.Store.SaveAccount(ctx, account), "SaveAccount must succeed")
+
+	mgr := agentnetwork.NewManager(am.Store, permissions.NewManager(am.Store), am, nil)
+
+	created, err := mgr.CreateBudgetRule(ctx, adminUserID, &agenttypes.AccountBudgetRule{
+		AccountID:    accountID,
+		Name:         "eng-monthly",
+		Enabled:      true,
+		TargetGroups: []string{"grp-eng"},
+		TargetUsers:  []string{"user-alice"},
+		Limits: agenttypes.PolicyLimits{
+			TokenLimit:  agenttypes.PolicyTokenLimit{Enabled: true, GroupCap: 100_000, UserCap: 10_000, WindowSeconds: 2_592_000},
+			BudgetLimit: agenttypes.PolicyBudgetLimit{Enabled: true, GroupCapUsd: 500, WindowSeconds: 2_592_000},
+		},
+	})
+	require.NoError(t, err, "CreateBudgetRule must succeed")
+	require.NotEmpty(t, created.ID, "create must mint an ID")
+
+	got, err := mgr.GetBudgetRule(ctx, accountID, adminUserID, created.ID)
+	require.NoError(t, err, "GetBudgetRule must succeed")
+	assert.Equal(t, "eng-monthly", got.Name, "name round-trips through the manager")
+	assert.Equal(t, []string{"grp-eng"}, got.TargetGroups, "target groups round-trip")
+	assert.Equal(t, int64(100_000), got.Limits.TokenLimit.GroupCap, "token group cap round-trips")
+
+	list, err := mgr.GetAllBudgetRules(ctx, accountID, adminUserID)
+	require.NoError(t, err, "GetAllBudgetRules must succeed")
+	require.Len(t, list, 1, "exactly the one created rule must be listed")
+
+	created.Limits.TokenLimit.GroupCap = 200_000
+	updated, err := mgr.UpdateBudgetRule(ctx, adminUserID, created)
+	require.NoError(t, err, "UpdateBudgetRule must succeed")
+	assert.Equal(t, int64(200_000), updated.Limits.TokenLimit.GroupCap, "updated cap must persist")
+
+	require.NoError(t, mgr.DeleteBudgetRule(ctx, accountID, adminUserID, created.ID), "DeleteBudgetRule must succeed")
+	_, err = mgr.GetBudgetRule(ctx, accountID, adminUserID, created.ID)
+	assert.Error(t, err, "get after delete must fail")
+}
+
+// TestAgentNetwork_UpdateSettings_PreservesImmutableAndTogglesCollection is the
+// GC-1 guard for UpdateSettings: it must apply the collection toggles while
+// preserving the immutable Domain/ProxyAddress assigned at bootstrap. The
+// request echoes the identity fields back — the PUT convention every other
+// endpoint follows — and a request echoing anything else is rejected outright
+// rather than quietly ignored.
+func TestAgentNetwork_UpdateSettings_PreservesImmutableAndTogglesCollection(t *testing.T) {
+	am, _, err := createManager(t)
+	require.NoError(t, err, "createManager must succeed")
+	ctx := context.Background()
+
+	const (
+		accountID   = "agent-net-settings-acct"
+		adminUserID = "agent-net-settings-admin"
+		clusterAddr = "eu.proxy.netbird.io"
+	)
+	account := newAccountWithId(ctx, accountID, adminUserID, "agent-net.test", "", "", false)
+	require.NoError(t, am.Store.SaveAccount(ctx, account), "SaveAccount must succeed")
+
+	mgr := agentnetwork.NewManager(am.Store, permissions.NewManager(am.Store), am, nil)
+
+	// Bootstrap is an explicit settings create; providers have no settings
+	// side effects anymore.
+	before, err := mgr.CreateSettings(ctx, adminUserID, agenttypes.DefaultSettings(accountID), clusterAddr, "")
+	require.NoError(t, err, "CreateSettings must bootstrap the row")
+	require.Equal(t, clusterAddr, before.ProxyAddress, "proxy address pinned at bootstrap")
+	require.NotEmpty(t, before.Domain, "endpoint allocated at bootstrap")
+	assert.False(t, before.EnablePromptCollection, "prompt collection defaults off")
+
+	_, err = mgr.CreateProvider(ctx, adminUserID, &agenttypes.Provider{
+		AccountID:  accountID,
+		ProviderID: "openai_api",
+		Name:       "openai",
+		// A private address: the save-time credential check leaves it
+		// unchecked rather than spending a dummy key against the real
+		// api.openai.com, which the vendor refuses and which would make
+		// this test depend on the runner having egress.
+		UpstreamURL: "https://10.255.255.1",
+		APIKey:      "sk-test",
+		Enabled:     true,
+		Models:      []agenttypes.ProviderModel{{ID: "gpt-5.4"}},
+	})
+	require.NoError(t, err, "CreateProvider must succeed")
+
+	// Flipping the toggles works when the request echoes the assigned
+	// identity. Retention is echoed too: UpdateSettings takes it verbatim, so
+	// omitting it would zero the account's retention.
+	updated, err := mgr.UpdateSettings(ctx, adminUserID, &agenttypes.Settings{
+		AccountID:              accountID,
+		Domain:                 before.Domain,
+		ProxyAddress:           before.ProxyAddress,
+		EnableLogCollection:    true,
+		EnablePromptCollection: true,
+		RedactPii:              true,
+		AccessLogRetentionDays: before.AccessLogRetentionDays,
+	})
+	require.NoError(t, err, "UpdateSettings must succeed")
+	assert.Equal(t, before.Domain, updated.Domain, "domain is immutable and must be preserved")
+	assert.Equal(t, before.ProxyAddress, updated.ProxyAddress, "proxy address is immutable and must be preserved")
+	assert.True(t, updated.EnableLogCollection, "log collection toggle must apply")
+	assert.True(t, updated.EnablePromptCollection, "prompt collection toggle must apply")
+	assert.True(t, updated.RedactPii, "redact toggle must apply")
+	assert.Equal(t, before.AccessLogRetentionDays, updated.AccessLogRetentionDays, "echoed retention must survive")
+
+	// Neither identity field can be smuggled into the row: a hand-rolled
+	// Settings value carrying a different endpoint or proxy address is
+	// rejected, not silently ignored.
+	for _, tc := range []struct {
+		name         string
+		domain       string
+		proxyAddress string
+	}{
+		{name: "foreign endpoint", domain: "evil.example.com", proxyAddress: before.ProxyAddress},
+		{name: "foreign proxy address", domain: before.Domain, proxyAddress: "attacker.cluster"},
+		{name: "empty identity echo", domain: "", proxyAddress: ""},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			_, err := mgr.UpdateSettings(ctx, adminUserID, &agenttypes.Settings{
+				AccountID:              accountID,
+				Domain:                 tc.domain,
+				ProxyAddress:           tc.proxyAddress,
+				EnableLogCollection:    false,
+				EnablePromptCollection: false,
+				RedactPii:              false,
+				AccessLogRetentionDays: before.AccessLogRetentionDays,
+			})
+			assert.Error(t, err, "a mismatched identity echo must be rejected")
+			assert.ErrorContains(t, err, "immutable", "the rejection must name the immutability rule")
+		})
+	}
+
+	// The rejected updates left the row exactly as the accepted one wrote it.
+	afterRejects, err := mgr.GetSettings(ctx, accountID, adminUserID)
+	require.NoError(t, err, "GetSettings must succeed")
+	assert.True(t, afterRejects.EnablePromptCollection, "a rejected update must not roll back the accepted toggles")
+
+	reloaded, err := am.Store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, accountID)
+	require.NoError(t, err)
+	assert.Equal(t, before.Domain, reloaded.Domain, "persisted domain unchanged")
+	assert.Equal(t, before.ProxyAddress, reloaded.ProxyAddress, "persisted proxy address unchanged")
+	assert.True(t, reloaded.EnablePromptCollection, "persisted prompt collection toggled on")
+}

management/server/agentnetwork_realstack_test.go

@@ -0,0 +1,224 @@
+package server
+
+import (
+	"context"
+	"net/netip"
+	"testing"
+	"time"
+
+	"github.com/miekg/dns"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	networkmap "github.com/netbirdio/netbird/management/internals/controllers/network_map"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
+	agenttypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	"github.com/netbirdio/netbird/management/internals/modules/permissions"
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+	nbproto "github.com/netbirdio/netbird/shared/management/proto"
+)
+
+// TestAgentNetwork_ProviderCRUD_FansOutToProxyAndClientPeers is the no-mock
+// integration test for the live propagation path: a provider/policy mutation
+// through the real agentnetwork.Manager triggers the real
+// DefaultAccountManager.UpdateAccountPeers, which runs the real network-map
+// controller (including AN-2b's injectAllProxyPolicies), and a network map is
+// computed and fanned out to BOTH the embedded proxy peer and the client peer.
+//
+// Unlike the synthesizer/reconcile unit tests, nothing here is mocked: real
+// SQLite store, real account manager + network-map controller, real
+// agentnetwork manager, real peer update channels. The client peer's delivered
+// map is asserted to actually carry the synth DNS surface, and provider
+// create/delete are exercised end to end.
+func TestAgentNetwork_ProviderCRUD_FansOutToProxyAndClientPeers(t *testing.T) {
+	am, updateManager, err := createManager(t)
+	require.NoError(t, err, "createManager must succeed")
+	ctx := context.Background()
+
+	const (
+		accountID    = "agent-net-acct-1"
+		adminUserID  = "agent-net-admin-1"
+		groupAID     = "agent-net-grp-A"
+		clusterAddr  = "eu.proxy.netbird.io"
+		clientKey    = "BhRPtynAAYRDy08+q4HTMsos8fs4plTP4NOSh7C1ry8="
+		proxyPeerID  = "agent-net-proxy-peer-1"
+		proxyPeerKey = "/yF0+vCfv+mRR5k0dca0TrGdO/oiNeAI58gToZm5NyI="
+		proxyIP      = "100.64.0.99"
+	)
+
+	account := newAccountWithId(ctx, accountID, adminUserID, "agent-net.test", "", "", false)
+	require.NoError(t, am.Store.SaveAccount(ctx, account), "SaveAccount must succeed")
+
+	// Real client peer through the production AddPeer path.
+	clientPeer := &nbpeer.Peer{
+		Key:      clientKey,
+		Name:     "agent-net-client",
+		DNSLabel: "agent-net-client",
+		Meta:     nbpeer.PeerSystemMeta{Hostname: "agent-net-client", GoOS: "linux", WtVersion: "development"},
+	}
+	addedClient, _, _, _, err := am.AddPeer(ctx, "", "", adminUserID, clientPeer, false)
+	require.NoError(t, err, "AddPeer must add the client peer")
+
+	// Inject a connected embedded proxy peer + put the client in the source group.
+	account, err = am.Store.GetAccount(ctx, accountID)
+	require.NoError(t, err)
+	account.Peers[proxyPeerID] = &nbpeer.Peer{
+		ID:        proxyPeerID,
+		AccountID: accountID,
+		Key:       proxyPeerKey,
+		IP:        netip.MustParseAddr(proxyIP),
+		Status:    &nbpeer.PeerStatus{Connected: true, LastSeen: time.Now().UTC()},
+		ProxyMeta: nbpeer.ProxyMeta{Embedded: true, Cluster: clusterAddr},
+		DNSLabel:  "agent-net-proxy",
+	}
+	account.Groups[groupAID] = &types.Group{ID: groupAID, Name: "groupA", Peers: []string{addedClient.ID}}
+	require.NoError(t, am.Store.SaveAccount(ctx, account), "SaveAccount must persist proxy peer + group")
+
+	// Subscribe to BOTH peers' update channels — this is how we observe the
+	// real fan-out.
+	clientCh := updateManager.CreateChannel(ctx, addedClient.ID)
+	proxyCh := updateManager.CreateChannel(ctx, proxyPeerID)
+	t.Cleanup(func() {
+		updateManager.CloseChannel(ctx, addedClient.ID)
+		updateManager.CloseChannel(ctx, proxyPeerID)
+	})
+	drain(clientCh)
+	drain(proxyCh)
+
+	// Real agentnetwork manager wired to the real account manager. proxyController
+	// is nil (no gRPC cluster fan-out here) — the reconcile still fires
+	// UpdateAccountPeers, which is the path under test.
+	agentMgr := agentnetwork.NewManager(am.Store, permissions.NewManager(am.Store), am, nil)
+
+	_, err = agentMgr.CreateSettings(ctx, adminUserID, agenttypes.DefaultSettings(accountID), clusterAddr, "")
+	require.NoError(t, err, "CreateSettings must bootstrap the endpoint")
+	// The bootstrap itself reconciles and queues updates on both channels;
+	// drain them so the fan-out assertions below can only be satisfied by the
+	// operation under test, not by this leftover.
+	drain(clientCh)
+	drain(proxyCh)
+
+	provider, err := agentMgr.CreateProvider(ctx, adminUserID, &agenttypes.Provider{
+		AccountID:  accountID,
+		ProviderID: "openai_api",
+		Name:       "openai-test",
+		// A private address: the save-time credential check leaves it
+		// unchecked rather than spending a dummy key against the real
+		// api.openai.com, which the vendor refuses and which would make
+		// this test depend on the runner having egress.
+		UpstreamURL: "https://10.255.255.1",
+		APIKey:      "sk-test-key",
+		Enabled:     true,
+		Models:      []agenttypes.ProviderModel{{ID: "gpt-5.4"}},
+	})
+	require.NoError(t, err, "CreateProvider must succeed")
+
+	policy, err := agentMgr.CreatePolicy(ctx, adminUserID, &agenttypes.Policy{
+		AccountID:              accountID,
+		Name:                   "p1",
+		Enabled:                true,
+		SourceGroups:           []string{groupAID},
+		DestinationProviderIDs: []string{provider.ID},
+	})
+	require.NoError(t, err, "CreatePolicy must succeed")
+
+	settings, err := am.Store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, accountID)
+	require.NoError(t, err)
+	fqdn := settings.Endpoint()
+
+	// Both peers must receive a fan-out. The provider-create reconcile fires
+	// before the policy exists (synth service then has no AccessGroups, so no
+	// zone), and the async update buffer can collapse/reorder updates — so we
+	// poll until the client's delivered map actually carries the synth record.
+	rdata := awaitZoneRData(clientCh, clusterAddr, fqdn, true)
+	assert.Equal(t, proxyIP, rdata,
+		"client peer's delivered network map must contain the synth DNS record pointing at the embedded proxy peer")
+	require.True(t, awaitUpdate(proxyCh), "embedded proxy peer must also receive a netmap update after create")
+
+	// UPDATE the provider — a new model on the existing service must still
+	// reconcile and keep the private surface routable (the live MODIFIED path).
+	provider.Models = append(provider.Models, agenttypes.ProviderModel{ID: "gpt-5.4-mini"})
+	_, err = agentMgr.UpdateProvider(ctx, adminUserID, provider)
+	require.NoError(t, err, "UpdateProvider must succeed")
+	assert.Equal(t, proxyIP, awaitZoneRData(clientCh, clusterAddr, fqdn, true),
+		"client peer must still resolve the synth record after the provider is updated")
+	require.True(t, awaitUpdate(proxyCh), "embedded proxy peer must also receive a netmap update after update")
+
+	// DELETE: detach the policy first (provider is in use), then drop the
+	// provider. Both peers update again and the synth surface disappears.
+	require.NoError(t, agentMgr.DeletePolicy(ctx, accountID, adminUserID, policy.ID), "DeletePolicy must succeed")
+	require.NoError(t, agentMgr.DeleteProvider(ctx, accountID, adminUserID, provider.ID), "DeleteProvider must succeed")
+
+	require.True(t, awaitUpdate(proxyCh), "embedded proxy peer must also receive a netmap update after delete")
+	assert.Empty(t, awaitZoneRData(clientCh, clusterAddr, fqdn, false),
+		"synth DNS record must be gone from the client's map after the provider is deleted")
+}
+
+// awaitZoneRData drains the channel for up to 8s. When wantPresent is true it
+// returns as soon as the synth record appears (its RData). When false it drains
+// to quiescence and returns the RData of the last delivered map (expected empty
+// once the provider is gone), tolerating stale buffered updates that still
+// carry the zone.
+func awaitZoneRData(ch <-chan *networkmap.UpdateMessage, clusterAddr, fqdn string, wantPresent bool) string {
+	deadline := time.After(8 * time.Second)
+	last := ""
+	for {
+		select {
+		case m := <-ch:
+			if m == nil {
+				continue
+			}
+			last = synthZoneRData(m.Update, clusterAddr, fqdn)
+			if wantPresent && last != "" {
+				return last
+			}
+		case <-time.After(750 * time.Millisecond):
+			return last
+		case <-deadline:
+			return last
+		}
+	}
+}
+
+// awaitUpdate reports whether at least one update arrives within the window.
+func awaitUpdate(ch <-chan *networkmap.UpdateMessage) bool {
+	select {
+	case m := <-ch:
+		return m != nil
+	case <-time.After(5 * time.Second):
+		return false
+	}
+}
+
+// drain empties any buffered updates (e.g. from AddPeer/SaveAccount) so the
+// next observation reflects the operation under test.
+func drain(ch <-chan *networkmap.UpdateMessage) {
+	for {
+		select {
+		case <-ch:
+		case <-time.After(200 * time.Millisecond):
+			return
+		}
+	}
+}
+
+// synthZoneRData returns the RData of the synth A record (record name == fqdn)
+// inside the cluster's custom zone, or "" when absent.
+func synthZoneRData(sync *nbproto.SyncResponse, clusterAddr, fqdn string) string {
+	if sync == nil {
+		return ""
+	}
+	for _, zone := range sync.GetNetworkMap().GetDNSConfig().GetCustomZones() {
+		if zone.GetDomain() != dns.Fqdn(clusterAddr) {
+			continue
+		}
+		for _, rec := range zone.GetRecords() {
+			if rec.GetName() == dns.Fqdn(fqdn) {
+				return rec.GetRData()
+			}
+		}
+	}
+	return ""
+}

management/server/auth/manager.go

@@ -12,6 +12,7 @@ import (
 	"github.com/netbirdio/netbird/shared/auth"
 
 	"github.com/netbirdio/netbird/base62"
+	"github.com/netbirdio/netbird/idp/dex"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 	nbjwt "github.com/netbirdio/netbird/shared/auth/jwt"
@@ -74,7 +75,10 @@ func (m *manager) ValidateAndParseToken(ctx context.Context, value string) (auth
 }
 
 func (m *manager) EnsureUserAccessByJWTGroups(ctx context.Context, userAuth auth.UserAuth, token *jwt.Token) (auth.UserAuth, error) {
-	if userAuth.IsChild || userAuth.IsPAT {
+	// Child accounts and PAT-authenticated requests do not use JWT group access checks.
+	// Embedded-Dex local users also skip them because local password authentication
+	// does not provide external IdP group claims.
+	if userAuth.IsChild || userAuth.IsPAT || dex.IsLocalUserID(userAuth.UserId) {
 		return userAuth, nil
 	}
 

management/server/auth/manager_test.go

@@ -16,6 +16,7 @@ import (
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
+	"github.com/netbirdio/netbird/idp/dex"
 	"github.com/netbirdio/netbird/management/server/auth"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
@@ -206,6 +207,43 @@ func TestAuthManager_EnsureUserAccessByJWTGroups(t *testing.T) {
 		_, err = manager.EnsureUserAccessByJWTGroups(context.Background(), userAuth, token)
 		require.Error(t, err, "ensure user access is not in allowed groups")
 	})
+
+	t.Run("Local embedded-Dex user is exempt from JWT allow-groups", func(t *testing.T) {
+		account.Settings.JWTGroupsEnabled = true
+		account.Settings.JWTGroupsClaimName = "idp-groups"
+		account.Settings.JWTAllowGroups = []string{"not-a-group"}
+		err := store.SaveAccount(context.Background(), account)
+		require.NoError(t, err, "save account failed")
+
+		// Local Dex users have a "local" connector encoded in their user ID.
+		localUserAuth := nbauth.UserAuth{
+			AccountId: account.Id,
+			Domain:    domain,
+			UserId:    dex.EncodeDexUserID("local-owner", "local"),
+		}
+
+		localUserAuth, err = manager.EnsureUserAccessByJWTGroups(context.Background(), localUserAuth, token)
+		require.NoError(t, err, "local user must not be locked out by JWT allow-groups (issue #5337)")
+		require.Len(t, localUserAuth.Groups, 0, "JWT groups must not be evaluated for local users")
+	})
+
+	t.Run("Federated embedded-Dex user is still subject to JWT allow-groups", func(t *testing.T) {
+		account.Settings.JWTGroupsEnabled = true
+		account.Settings.JWTGroupsClaimName = "idp-groups"
+		account.Settings.JWTAllowGroups = []string{"not-a-group"}
+		err := store.SaveAccount(context.Background(), account)
+		require.NoError(t, err, "save account failed")
+
+		// A federated user (non-"local" connector) must remain restricted.
+		fedUserAuth := nbauth.UserAuth{
+			AccountId: account.Id,
+			Domain:    domain,
+			UserId:    dex.EncodeDexUserID("entra-user", "entra"),
+		}
+
+		_, err = manager.EnsureUserAccessByJWTGroups(context.Background(), fedUserAuth, token)
+		require.Error(t, err, "federated user must still be restricted by JWT allow-groups")
+	})
 }
 
 func TestAuthManager_ValidateAndParseToken(t *testing.T) {

management/server/auth/session.go

@@ -0,0 +1,57 @@
+package auth
+
+import (
+	"context"
+	"crypto/sha256"
+	"encoding/hex"
+	"errors"
+	"fmt"
+	"time"
+)
+
+const (
+	usedTokenKeyPrefix = "jwt-used:"
+	usedTokenMarker    = "1"
+)
+
+var (
+	ErrTokenAlreadyUsed = errors.New("JWT already used")
+	ErrTokenExpired     = errors.New("JWT expired")
+)
+
+// TokenCache atomically records used JWTs until their expiration.
+type TokenCache interface {
+	SetNX(ctx context.Context, key, value string, ttl time.Duration) (bool, error)
+}
+
+type SessionStore struct {
+	cache TokenCache
+}
+
+func NewSessionStore(cacheStore TokenCache) *SessionStore {
+	return &SessionStore{cache: cacheStore}
+}
+
+// RegisterToken records a JWT until its exp time and rejects reuse.
+func (s *SessionStore) RegisterToken(ctx context.Context, token string, expiresAt time.Time) error {
+	ttl := time.Until(expiresAt)
+	if ttl <= 0 {
+		return ErrTokenExpired
+	}
+
+	key := usedTokenKeyPrefix + hashToken(token)
+	created, err := s.cache.SetNX(ctx, key, usedTokenMarker, ttl)
+	if err != nil {
+		return fmt.Errorf("store used token entry: %w", err)
+	}
+	if !created {
+		return ErrTokenAlreadyUsed
+	}
+
+	return nil
+}
+
+func hashToken(token string) string {
+	sum := sha256.Sum256([]byte(token))
+	return hex.EncodeToString(sum[:])
+}

management/server/auth/session_test.go

@@ -0,0 +1,133 @@
+package auth
+
+import (
+	"context"
+	"errors"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	nbcache "github.com/netbirdio/netbird/management/server/cache"
+)
+
+func newTestSessionStore(t *testing.T) *SessionStore {
+	t.Helper()
+	cacheStore, err := nbcache.NewStore(context.Background(), time.Hour, time.Hour, 100)
+	require.NoError(t, err)
+	return NewSessionStore(cacheStore)
+}
+
+func TestSessionStore_FirstRegisterSucceeds(t *testing.T) {
+	s := newTestSessionStore(t)
+	ctx := context.Background()
+
+	require.NoError(t, s.RegisterToken(ctx, "token", time.Now().Add(time.Hour)))
+}
+
+func TestSessionStore_RegisterSameTokenTwiceIsRejected(t *testing.T) {
+	s := newTestSessionStore(t)
+	ctx := context.Background()
+	token := "token"
+	exp := time.Now().Add(time.Hour)
+
+	require.NoError(t, s.RegisterToken(ctx, token, exp))
+
+	err := s.RegisterToken(ctx, token, exp)
+	require.Error(t, err)
+	assert.ErrorIs(t, err, ErrTokenAlreadyUsed)
+}
+
+func TestSessionStore_ConcurrentRegistrationAllowsOneCaller(t *testing.T) {
+	s := newTestSessionStore(t)
+	ctx := context.Background()
+	const attempts = 100
+
+	start := make(chan struct{})
+	results := make(chan error, attempts)
+	for range attempts {
+		go func() {
+			<-start
+			results <- s.RegisterToken(ctx, "token", time.Now().Add(time.Hour))
+		}()
+	}
+	close(start)
+
+	succeeded := 0
+	alreadyUsed := 0
+	for range attempts {
+		err := <-results
+		switch {
+		case err == nil:
+			succeeded++
+		case errors.Is(err, ErrTokenAlreadyUsed):
+			alreadyUsed++
+		default:
+			require.NoError(t, err, "concurrent registration returned an unexpected error")
+		}
+	}
+
+	assert.Equal(t, 1, succeeded, "exactly one concurrent caller should register the token")
+	assert.Equal(t, attempts-1, alreadyUsed, "every other caller should be rejected as already used")
+}
+
+func TestSessionStore_RegisterDifferentTokensAreIndependent(t *testing.T) {
+	s := newTestSessionStore(t)
+	ctx := context.Background()
+	exp := time.Now().Add(time.Hour)
+
+	require.NoError(t, s.RegisterToken(ctx, "tokenA", exp))
+	require.NoError(t, s.RegisterToken(ctx, "tokenB", exp))
+}
+
+func TestSessionStore_RegisterWithPastExpiryIsRejected(t *testing.T) {
+	s := newTestSessionStore(t)
+	ctx := context.Background()
+	token := "token"
+
+	err := s.RegisterToken(ctx, token, time.Now().Add(-time.Second))
+	require.Error(t, err)
+	assert.ErrorIs(t, err, ErrTokenExpired)
+}
+
+func TestSessionStore_EntryEvictsAtTTLAndAllowsReRegistration(t *testing.T) {
+	s := newTestSessionStore(t)
+	ctx := context.Background()
+	token := "token"
+
+	require.NoError(t, s.RegisterToken(ctx, token, time.Now().Add(50*time.Millisecond)))
+
+	err := s.RegisterToken(ctx, token, time.Now().Add(50*time.Millisecond))
+	assert.ErrorIs(t, err, ErrTokenAlreadyUsed)
+
+	time.Sleep(120 * time.Millisecond)
+
+	require.NoError(t, s.RegisterToken(ctx, token, time.Now().Add(time.Hour)))
+}
+
+type failingTokenCache struct {
+	err error
+}
+
+func (f failingTokenCache) SetNX(context.Context, string, string, time.Duration) (bool, error) {
+	return false, f.err
+}
+
+func TestSessionStore_CacheErrorIsReturned(t *testing.T) {
+	cacheErr := errors.New("cache unavailable")
+	s := NewSessionStore(failingTokenCache{err: cacheErr})
+
+	err := s.RegisterToken(context.Background(), "token", time.Now().Add(time.Hour))
+	require.Error(t, err, "cache failure should be surfaced to the caller")
+	assert.ErrorIs(t, err, cacheErr, "cache error should be wrapped, not replaced")
+}
+
+func TestHashToken_StableAndDoesNotLeak(t *testing.T) {
+	a := hashToken("tokenA")
+	b := hashToken("tokenB")
+	assert.Equal(t, a, hashToken("tokenA"), "hash must be deterministic")
+	assert.NotEqual(t, a, b, "different tokens must hash differently")
+	assert.Len(t, a, 64, "sha256 hex must be 64 chars")
+	assert.NotContains(t, a, "tokenA", "raw token must not appear in hash")
+}

management/server/cache/memory.go

@@ -0,0 +1,57 @@
+package cache
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"time"
+
+	"github.com/eko/gocache/lib/v4/store"
+	gocachestore "github.com/eko/gocache/store/go_cache/v4"
+	gocache "github.com/patrickmn/go-cache"
+)
+
+type goCacheStore struct {
+	store.StoreInterface
+	client *gocache.Cache
+	mu     sync.Mutex
+}
+
+func newMemoryStore(maxTimeout, cleanupInterval time.Duration) Store {
+	client := gocache.New(maxTimeout, cleanupInterval)
+	return &goCacheStore{
+		StoreInterface: gocachestore.NewGoCache(client),
+		client:         client,
+	}
+}
+
+func (s *goCacheStore) SetNX(_ context.Context, key, value string, ttl time.Duration) (bool, error) {
+	// Add only returns an error when a non-expired entry already exists.
+	if err := s.client.Add(key, value, ttl); err != nil {
+		return false, nil //nolint:nilerr
+	}
+	return true, nil
+}
+
+// GetDel reads the value under key and removes it. go-cache has no native read-and-delete
+// and releases its own lock between the two calls, so mu holds the pair together and no
+// value is consumed twice.
+//
+// Writes do not take mu: a Set landing mid-pair is lost, since GetDel returns the prior
+// value and deletes the new one. Callers must write a consumed key only once.
+func (s *goCacheStore) GetDel(_ context.Context, key string) (string, bool, error) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	value, found := s.client.Get(key)
+	if !found {
+		return "", false, nil
+	}
+	s.client.Delete(key)
+
+	str, ok := value.(string)
+	if !ok {
+		return "", false, fmt.Errorf("cached value is %T, not a string", value)
+	}
+	return str, true, nil
+}

management/server/cache/memory_test.go

@@ -0,0 +1,76 @@
+package cache_test
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/server/cache"
+)
+
+func TestMemoryStore(t *testing.T) {
+	memStore, err := cache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
+	require.NoError(t, err, "couldn't create memory store")
+
+	ctx := context.Background()
+	key, value := "testing", "tested"
+	err = memStore.Set(ctx, key, value)
+	assert.NoError(t, err, "couldn't set testing data")
+
+	result, err := memStore.Get(ctx, key)
+	assert.NoError(t, err, "couldn't get testing data")
+	assert.Equal(t, value, result, "value returned doesn't match testing data")
+
+	created, err := memStore.SetNX(ctx, "conditional", value, 100*time.Millisecond)
+	require.NoError(t, err, "couldn't conditionally set testing data")
+	require.True(t, created, "first conditional set should create the entry")
+
+	created, err = memStore.SetNX(ctx, "conditional", value, 100*time.Millisecond)
+	require.NoError(t, err, "couldn't conditionally check testing data")
+	require.False(t, created, "second conditional set should not replace the entry")
+
+	// test expiration
+	time.Sleep(300 * time.Millisecond)
+	_, err = memStore.Get(ctx, key)
+	assert.Error(t, err, "value should not be found")
+}
+
+func TestMemoryStoreGetDel(t *testing.T) {
+	ctx := context.Background()
+	newStore := func(t *testing.T) cache.Store {
+		t.Helper()
+		memStore, err := cache.NewStore(ctx, time.Minute, time.Minute, 100)
+		require.NoError(t, err, "couldn't create memory store")
+
+		return memStore
+	}
+
+	const (
+		key   = "consume"
+		value = "verifier"
+	)
+
+	t.Run("exactly one concurrent caller consumes the key", func(t *testing.T) {
+		memStore := newStore(t)
+		require.NoError(t, memStore.Set(ctx, key, value), "couldn't set testing data")
+
+		assertGetDelConsumedOnce(ctx, t, []cache.Store{memStore}, key, value)
+		assertGetDelMisses(ctx, t, memStore, key)
+	})
+
+	t.Run("missing key is not an error", func(t *testing.T) {
+		assertGetDelMisses(ctx, t, newStore(t), "never-set")
+	})
+
+	t.Run("expired key is not found", func(t *testing.T) {
+		memStore := newStore(t)
+		_, err := memStore.SetNX(ctx, key, value, 50*time.Millisecond)
+		require.NoError(t, err, "couldn't set testing data")
+
+		time.Sleep(100 * time.Millisecond)
+		assertGetDelMisses(ctx, t, memStore, key)
+	})
+}

management/server/cache/redis.go

@@ -0,0 +1,63 @@
+package cache
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"math"
+	"time"
+
+	"github.com/eko/gocache/lib/v4/store"
+	redisstore "github.com/eko/gocache/store/redis/v4"
+	"github.com/redis/go-redis/v9"
+	log "github.com/sirupsen/logrus"
+)
+
+type redisStore struct {
+	store.StoreInterface
+	client *redis.Client
+}
+
+func getRedisStore(ctx context.Context, redisEnvAddr string, maxConn int) (Store, error) {
+	options, err := redis.ParseURL(redisEnvAddr)
+	if err != nil {
+		return nil, fmt.Errorf("parsing redis cache url: %s", err)
+	}
+
+	options.MaxIdleConns = int(math.Ceil(float64(maxConn) * 0.5)) // 50% of max conns
+	options.MinIdleConns = int(math.Ceil(float64(maxConn) * 0.1)) // 10% of max conns
+	options.MaxActiveConns = maxConn
+	options.ConnMaxIdleTime = 30 * time.Minute
+	options.ConnMaxLifetime = 0
+	options.PoolTimeout = 10 * time.Second
+	redisClient := redis.NewClient(options)
+	subCtx, cancel := context.WithTimeout(ctx, 2*time.Second)
+	defer cancel()
+
+	_, err = redisClient.Ping(subCtx).Result()
+	if err != nil {
+		return nil, err
+	}
+
+	log.WithContext(subCtx).Infof("using redis cache at %s", redisEnvAddr)
+
+	return &redisStore{
+		StoreInterface: redisstore.NewRedis(redisClient),
+		client:         redisClient,
+	}, nil
+}
+
+func (s *redisStore) SetNX(ctx context.Context, key, value string, ttl time.Duration) (bool, error) {
+	return s.client.SetNX(ctx, key, value, ttl).Result()
+}
+
+func (s *redisStore) GetDel(ctx context.Context, key string) (string, bool, error) {
+	value, err := s.client.GetDel(ctx, key).Result()
+	if errors.Is(err, redis.Nil) {
+		return "", false, nil
+	}
+	if err != nil {
+		return "", false, err
+	}
+	return value, true, nil
+}

management/server/cache/redis_test.go

@@ -0,0 +1,153 @@
+package cache_test
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/eko/gocache/lib/v4/store"
+	"github.com/redis/go-redis/v9"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	testcontainersredis "github.com/testcontainers/testcontainers-go/modules/redis"
+
+	"github.com/netbirdio/netbird/management/server/cache"
+)
+
+func startRedis(t *testing.T) string {
+	t.Helper()
+
+	ctx := context.Background()
+	redisContainer, err := testcontainersredis.Run(ctx, "redis:7")
+	require.NoError(t, err, "couldn't start redis container")
+
+	t.Cleanup(func() {
+		if err := redisContainer.Terminate(ctx); err != nil {
+			t.Logf("failed to terminate container: %s", err)
+		}
+	})
+
+	redisURL, err := redisContainer.ConnectionString(ctx)
+	require.NoError(t, err, "couldn't get connection string")
+
+	t.Setenv(cache.RedisStoreEnvVar, redisURL)
+	return redisURL
+}
+
+func newRedisStore(t *testing.T) cache.Store {
+	t.Helper()
+
+	redisStore, err := cache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
+	require.NoError(t, err)
+
+	return redisStore
+}
+
+func TestRedisStoreConnectionFailure(t *testing.T) {
+	t.Setenv(cache.RedisStoreEnvVar, "redis://127.0.0.1:6379")
+	_, err := cache.NewStore(context.Background(), 10*time.Millisecond, 30*time.Millisecond, 100)
+	require.Error(t, err, "getting redis cache store should return error")
+}
+
+func TestRedisStoreConnectionSuccess(t *testing.T) {
+	ctx := context.Background()
+	redisURL := startRedis(t)
+	redisStore := newRedisStore(t)
+
+	key, value := "testing", "tested"
+	err := redisStore.Set(ctx, key, value, store.WithExpiration(100*time.Millisecond))
+	assert.NoError(t, err, "couldn't set testing data")
+
+	result, err := redisStore.Get(ctx, key)
+	assert.NoError(t, err, "couldn't get testing data")
+	assert.Equal(t, value, result, "value returned doesn't match testing data")
+
+	options, err := redis.ParseURL(redisURL)
+	require.NoError(t, err, "parsing redis cache url")
+
+	redisClient := redis.NewClient(options)
+	r, err := redisClient.Get(ctx, key).Result()
+	assert.NoError(t, err, "couldn't get testing data from redis")
+	assert.Equal(t, value, r, "value returned from redis doesn't match testing data")
+
+	// test expiration
+	time.Sleep(300 * time.Millisecond)
+	_, err = redisStore.Get(ctx, key)
+	assert.Error(t, err, "value should not be found")
+}
+
+func TestRedisStoreSetNX(t *testing.T) {
+	ctx := context.Background()
+	redisURL := startRedis(t)
+	redisStore, secondRedisStore := newRedisStore(t), newRedisStore(t)
+
+	const (
+		key   = "conditional"
+		value = "tested"
+	)
+
+	start := make(chan struct{})
+	type setResult struct {
+		created bool
+		err     error
+	}
+	results := make(chan setResult, 2)
+	for _, cacheStore := range []cache.Store{redisStore, secondRedisStore} {
+		go func() {
+			<-start
+			created, err := cacheStore.SetNX(ctx, key, value, time.Minute)
+			results <- setResult{created: created, err: err}
+		}()
+	}
+	close(start)
+
+	created := 0
+	for range 2 {
+		result := <-results
+		require.NoError(t, result.err, "conditional redis set failed")
+		if result.created {
+			created++
+		}
+	}
+	require.Equal(t, 1, created, "expected exactly one redis client to create the entry")
+
+	options, err := redis.ParseURL(redisURL)
+	require.NoError(t, err, "parsing redis cache url")
+
+	ttl, err := redis.NewClient(options).PTTL(ctx, key).Result()
+	require.NoError(t, err, "couldn't read entry TTL")
+	require.Positive(t, ttl, "created entry should have a positive TTL")
+}
+
+func TestRedisStoreGetDel(t *testing.T) {
+	ctx := context.Background()
+	startRedis(t)
+	redisStore, secondRedisStore := newRedisStore(t), newRedisStore(t)
+
+	const (
+		key   = "consume"
+		value = "verifier"
+	)
+
+	t.Run("exactly one caller across independent clients consumes the key", func(t *testing.T) {
+		// A generous TTL: the key is consumed explicitly, so expiry racing the
+		// concurrent callers would only make the test flaky on a loaded runner.
+		err := redisStore.Set(ctx, key, value, store.WithExpiration(time.Minute))
+		require.NoError(t, err, "couldn't set value to consume")
+
+		assertGetDelConsumedOnce(ctx, t, []cache.Store{redisStore, secondRedisStore}, key, value)
+		assertGetDelMisses(ctx, t, secondRedisStore, key)
+	})
+
+	t.Run("missing key is not an error", func(t *testing.T) {
+		assertGetDelMisses(ctx, t, redisStore, "never-set")
+	})
+
+	t.Run("expired key is not found", func(t *testing.T) {
+		err := redisStore.Set(ctx, key, value, store.WithExpiration(50*time.Millisecond))
+		require.NoError(t, err, "couldn't set value to consume")
+
+		time.Sleep(100 * time.Millisecond)
+		assertGetDelMisses(ctx, t, redisStore, key)
+	})
+}

management/server/cache/store.go

@@ -2,17 +2,10 @@ package cache
 
 import (
 	"context"
-	"fmt"
-	"math"
 	"os"
 	"time"
 
 	"github.com/eko/gocache/lib/v4/store"
-	gocache_store "github.com/eko/gocache/store/go_cache/v4"
-	redis_store "github.com/eko/gocache/store/redis/v4"
-	gocache "github.com/patrickmn/go-cache"
-	"github.com/redis/go-redis/v9"
-	log "github.com/sirupsen/logrus"
 )
 
 // RedisStoreEnvVar is the environment variable that determines if a redis store should be used.
@@ -31,15 +24,23 @@ const (
 	DefaultStoreMaxConn = 1000
 )
 
+// Store extends the shared cache interface with conditional and consuming operations.
+type Store interface {
+	store.StoreInterface
+	// SetNX stores a value with a TTL only when the key does not exist.
+	SetNX(ctx context.Context, key, value string, ttl time.Duration) (bool, error)
+	// GetDel reads a value and removes it, so only one caller can consume a key.
+	GetDel(ctx context.Context, key string) (value string, found bool, err error)
+}
+
 // NewStore creates a new cache store with the given max timeout and cleanup interval. It checks for the environment Variable RedisStoreEnvVar
 // to determine if a redis store should be used. If the environment variable is set, it will attempt to connect to the redis store.
-func NewStore(ctx context.Context, maxTimeout, cleanupInterval time.Duration, maxConn int) (store.StoreInterface, error) {
+func NewStore(ctx context.Context, maxTimeout, cleanupInterval time.Duration, maxConn int) (Store, error) {
 	redisAddr := GetAddrFromEnv()
 	if redisAddr != "" {
 		return getRedisStore(ctx, redisAddr, maxConn)
 	}
-	goc := gocache.New(maxTimeout, cleanupInterval)
-	return gocache_store.NewGoCache(goc), nil
+	return newMemoryStore(maxTimeout, cleanupInterval), nil
 }
 
 // GetAddrFromEnv returns the redis address from the environment variable RedisStoreEnvVar or its legacy counterpart.
@@ -50,29 +51,3 @@ func GetAddrFromEnv() string {
 	}
 	return addr
 }
-
-func getRedisStore(ctx context.Context, redisEnvAddr string, maxConn int) (store.StoreInterface, error) {
-	options, err := redis.ParseURL(redisEnvAddr)
-	if err != nil {
-		return nil, fmt.Errorf("parsing redis cache url: %s", err)
-	}
-
-	options.MaxIdleConns = int(math.Ceil(float64(maxConn) * 0.5)) // 50% of max conns
-	options.MinIdleConns = int(math.Ceil(float64(maxConn) * 0.1)) // 10% of max conns
-	options.MaxActiveConns = maxConn
-	options.ConnMaxIdleTime = 30 * time.Minute
-	options.ConnMaxLifetime = 0
-	options.PoolTimeout = 10 * time.Second
-	redisClient := redis.NewClient(options)
-	subCtx, cancel := context.WithTimeout(ctx, 2*time.Second)
-	defer cancel()
-
-	_, err = redisClient.Ping(subCtx).Result()
-	if err != nil {
-		return nil, err
-	}
-
-	log.WithContext(subCtx).Infof("using redis cache at %s", redisEnvAddr)
-
-	return redis_store.NewRedis(redisClient), nil
-}

management/server/cache/store_test.go

@@ -3,101 +3,53 @@ package cache_test
 import (
 	"context"
 	"testing"
-	"time"
 
-	"github.com/eko/gocache/lib/v4/store"
-	"github.com/redis/go-redis/v9"
-	testcontainersredis "github.com/testcontainers/testcontainers-go/modules/redis"
+	"github.com/stretchr/testify/require"
 
 	"github.com/netbirdio/netbird/management/server/cache"
 )
 
-func TestMemoryStore(t *testing.T) {
-	memStore, err := cache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
-	if err != nil {
-		t.Fatalf("couldn't create memory store: %s", err)
-	}
-	ctx := context.Background()
-	key, value := "testing", "tested"
-	err = memStore.Set(ctx, key, value)
-	if err != nil {
-		t.Errorf("couldn't set testing data: %s", err)
-	}
-	result, err := memStore.Get(ctx, key)
-	if err != nil {
-		t.Errorf("couldn't get testing data: %s", err)
-	}
-	if value != result.(string) {
-		t.Errorf("value returned doesn't match testing data, got %s, expected %s", result, value)
-	}
-	// test expiration
-	time.Sleep(300 * time.Millisecond)
-	_, err = memStore.Get(ctx, key)
-	if err == nil {
-		t.Error("value should not be found")
-	}
-}
+func assertGetDelConsumedOnce(ctx context.Context, t *testing.T, stores []cache.Store, key, value string) {
+	t.Helper()
 
-func TestRedisStoreConnectionFailure(t *testing.T) {
-	t.Setenv(cache.RedisStoreEnvVar, "redis://127.0.0.1:6379")
-	_, err := cache.NewStore(context.Background(), 10*time.Millisecond, 30*time.Millisecond, 100)
-	if err == nil {
-		t.Fatal("getting redis cache store should return error")
-	}
-}
+	const getDelAttempts = 64
 
-func TestRedisStoreConnectionSuccess(t *testing.T) {
-	ctx := context.Background()
-	redisContainer, err := testcontainersredis.Run(ctx, "redis:7")
-	if err != nil {
-		t.Fatalf("couldn't start redis container: %s", err)
+	type getDelResult struct {
+		value string
+		found bool
+		err   error
 	}
-	defer func() {
-		if err := redisContainer.Terminate(ctx); err != nil {
-			t.Logf("failed to terminate container: %s", err)
+
+	start := make(chan struct{})
+	results := make(chan getDelResult, getDelAttempts)
+	for i := range getDelAttempts {
+		cacheStore := stores[i%len(stores)]
+		go func() {
+			<-start
+			value, found, err := cacheStore.GetDel(ctx, key)
+			results <- getDelResult{value: value, found: found, err: err}
+		}()
+	}
+	close(start)
+
+	consumers := 0
+	for range getDelAttempts {
+		result := <-results
+		require.NoError(t, result.err, "concurrent GetDel failed")
+		if !result.found {
+			continue
 		}
-	}()
-	redisURL, err := redisContainer.ConnectionString(ctx)
-	if err != nil {
-		t.Fatalf("couldn't get connection string: %s", err)
-	}
-
-	t.Setenv(cache.RedisStoreEnvVar, redisURL)
-	redisStore, err := cache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
-	if err != nil {
-		t.Fatalf("couldn't create redis store: %s", err)
-	}
-
-	key, value := "testing", "tested"
-	err = redisStore.Set(ctx, key, value, store.WithExpiration(100*time.Millisecond))
-	if err != nil {
-		t.Errorf("couldn't set testing data: %s", err)
-	}
-	result, err := redisStore.Get(ctx, key)
-	if err != nil {
-		t.Errorf("couldn't get testing data: %s", err)
-	}
-	if value != result.(string) {
-		t.Errorf("value returned doesn't match testing data, got %s, expected %s", result, value)
-	}
-
-	options, err := redis.ParseURL(redisURL)
-	if err != nil {
-		t.Errorf("parsing redis cache url: %s", err)
-	}
-
-	redisClient := redis.NewClient(options)
-	r, e := redisClient.Get(ctx, key).Result()
-	if e != nil {
-		t.Errorf("couldn't get testing data from redis: %s", e)
-	}
-	if value != r {
-		t.Errorf("value returned from redis doesn't match testing data, got %s, expected %s", r, value)
-	}
-	// test expiration
-	time.Sleep(300 * time.Millisecond)
-	_, err = redisStore.Get(ctx, key)
-	if err == nil {
-		t.Error("value should not be found")
+		consumers++
+		require.Equal(t, value, result.value, "consumed value doesn't match testing data")
 	}
+	require.Equal(t, 1, consumers, "expected exactly one consumer")
+}
+
+func assertGetDelMisses(ctx context.Context, t *testing.T, cacheStore cache.Store, key string) {
+	t.Helper()
+
+	value, found, err := cacheStore.GetDel(ctx, key)
+	require.NoError(t, err, "GetDel on a missing key should not error")
+	require.False(t, found, "GetDel should not find key %q, got value %q", key, value)
+	require.Empty(t, value, "GetDel should return an empty value when not found")
 }

management/server/context/keys.go

@@ -1,10 +1,28 @@
 package context
 
-import "github.com/netbirdio/netbird/shared/context"
+import (
+	"context"
+
+	nbcontext "github.com/netbirdio/netbird/shared/context"
+)
 
 const (
-	RequestIDKey = context.RequestIDKey
-	AccountIDKey = context.AccountIDKey
-	UserIDKey    = context.UserIDKey
-	PeerIDKey    = context.PeerIDKey
+	RequestIDKey = nbcontext.RequestIDKey
+	AccountIDKey = nbcontext.AccountIDKey
+	RoleKey      = nbcontext.RoleKey
+	UserIDKey    = nbcontext.UserIDKey
+	PeerIDKey    = nbcontext.PeerIDKey
+	UserAgentKey = nbcontext.UserAgentKey
 )
+
+// RoleFromContext returns the role stored in ctx, or empty string and false if absent.
+func RoleFromContext(ctx context.Context) (string, bool) {
+	role, ok := ctx.Value(RoleKey).(string)
+	return role, ok
+}
+
+// WithRole returns a new context carrying the given role.
+func WithRole(ctx context.Context, role string) context.Context {
+	//nolint
+	return context.WithValue(ctx, RoleKey, role)
+}

management/server/dns.go

@@ -8,6 +8,7 @@ import (
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/management/server/util"
@@ -29,8 +30,9 @@ func (am *DefaultAccountManager) SaveDNSSettings(ctx context.Context, accountID
 		return status.Errorf(status.InvalidArgument, "the dns settings provided are nil")
 	}
 
-	var updateAccountPeers bool
 	var eventsToStore []func()
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		var err error
@@ -46,11 +48,6 @@ func (am *DefaultAccountManager) SaveDNSSettings(ctx context.Context, accountID
 		addedGroups := util.Difference(dnsSettingsToSave.DisabledManagementGroups, oldSettings.DisabledManagementGroups)
 		removedGroups := util.Difference(oldSettings.DisabledManagementGroups, dnsSettingsToSave.DisabledManagementGroups)
 
-		updateAccountPeers, err = areDNSSettingChangesAffectPeers(ctx, transaction, accountID, addedGroups, removedGroups)
-		if err != nil {
-			return err
-		}
-
 		events := am.prepareDNSSettingsEvents(ctx, transaction, accountID, userID, addedGroups, removedGroups)
 		eventsToStore = append(eventsToStore, events...)
 
@@ -58,6 +55,11 @@ func (am *DefaultAccountManager) SaveDNSSettings(ctx context.Context, accountID
 			return err
 		}
 
+		change = affectedpeers.Change{DistributionGroupIDs: slices.Concat(addedGroups, removedGroups)}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
+			return err
+		}
+
 		return transaction.IncrementNetworkSerial(ctx, accountID)
 	})
 	if err != nil {
@@ -68,9 +70,7 @@ func (am *DefaultAccountManager) SaveDNSSettings(ctx context.Context, accountID
 		storeEvent()
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -116,20 +116,6 @@ func (am *DefaultAccountManager) prepareDNSSettingsEvents(ctx context.Context, t
 	return eventsToStore
 }
 
-// areDNSSettingChangesAffectPeers checks if the DNS settings changes affect any peers.
-func areDNSSettingChangesAffectPeers(ctx context.Context, transaction store.Store, accountID string, addedGroups, removedGroups []string) (bool, error) {
-	hasPeers, err := anyGroupHasPeersOrResources(ctx, transaction, accountID, addedGroups)
-	if err != nil {
-		return false, err
-	}
-
-	if hasPeers {
-		return true, nil
-	}
-
-	return anyGroupHasPeersOrResources(ctx, transaction, accountID, removedGroups)
-}
-
 // validateDNSSettings validates the DNS settings.
 func validateDNSSettings(ctx context.Context, transaction store.Store, accountID string, settings *types.DNSSettings) error {
 	if len(settings.DisabledManagementGroups) == 0 {

management/server/dns_test.go

@@ -6,8 +6,8 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
 	"github.com/stretchr/testify/assert"
+	"go.uber.org/mock/gomock"
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
@@ -223,7 +223,7 @@ func createDNSManager(t *testing.T) (*DefaultAccountManager, error) {
 
 	updateManager := update_channel.NewPeersUpdateManager(metrics)
 	requestBuffer := NewAccountRequestBuffer(ctx, store)
-	networkMapController := controller.NewController(ctx, store, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.test", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peers.NewManager(store)), &config.Config{})
+	networkMapController := controller.NewController(ctx, store, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.test", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peers.NewManager(store)), &config.Config{}, nil)
 
 	return BuildManager(context.Background(), nil, store, networkMapController, job.NewJobManager(nil, store, peersManager), nil, "", eventStore, nil, false, MockIntegratedValidator{}, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
 }
@@ -287,11 +287,11 @@ func initTestDNSAccount(t *testing.T, am *DefaultAccountManager) (*types.Account
 		return nil, err
 	}
 
-	savedPeer1, _, _, err := am.AddPeer(context.Background(), "", "", dnsAdminUserID, peer1, false)
+	savedPeer1, _, _, _, err := am.AddPeer(context.Background(), "", "", dnsAdminUserID, peer1, false)
 	if err != nil {
 		return nil, err
 	}
-	_, _, _, err = am.AddPeer(context.Background(), "", "", dnsAdminUserID, peer2, false)
+	_, _, _, _, err = am.AddPeer(context.Background(), "", "", dnsAdminUserID, peer2, false)
 	if err != nil {
 		return nil, err
 	}
@@ -447,7 +447,7 @@ func TestDNSAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -467,7 +467,7 @@ func TestDNSAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -507,7 +507,7 @@ func TestDNSAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})

management/server/geolocation/database.go

@@ -3,6 +3,7 @@ package geolocation
 import (
 	"context"
 	"encoding/csv"
+	"fmt"
 	"io"
 	"os"
 	"path"
@@ -21,6 +22,8 @@ const (
 	geoLiteCitySha256ZipURL = "https://pkgs.netbird.io/geolocation-dbs/GeoLite2-City-CSV/download?suffix=zip.sha256"
 	geoLiteCityMMDB         = "GeoLite2-City.mmdb"
 	geoLiteCityCSV          = "GeoLite2-City-Locations-en.csv"
+
+	geonamesCsvFields = 14
 )
 
 // loadGeolocationDatabases loads the MaxMind databases.
@@ -160,6 +163,10 @@ func loadGeonamesCsv(filepath string) ([]GeoNames, error) {
 		if index == 0 {
 			continue
 		}
+		if len(record) < geonamesCsvFields {
+			return nil, fmt.Errorf("geonames csv record %d has %d fields, want at least %d", index, len(record), geonamesCsvFields)
+		}
+
 		geoNameID, err := strconv.Atoi(record[0])
 		if err != nil {
 			return nil, err

management/server/geolocation/geolocation.go

@@ -242,7 +242,11 @@ func getDatabaseFilename(ctx context.Context, databaseURL string, filenamePatter
 	// strip suffixes that may be nested, such as .tar.gz
 	basename := strings.SplitN(filename, ".", 2)[0]
 	// get date version from basename
-	date := strings.SplitN(basename, "_", 2)[1]
+	parts := strings.SplitN(basename, "_", 2)
+	if len(parts) < 2 || parts[1] == "" {
+		return "", fmt.Errorf("unexpected database filename %q: missing date suffix", filename)
+	}
+	date := parts[1]
 	// format db as "GeoLite2-Cities-{maxmind|geonames}_{DATE}.{mmdb|db}"
 	databaseFilename := filepath.Base(strings.Replace(filenamePattern, "*", date, 1))
 

management/server/geolocation/utils.go

@@ -184,7 +184,12 @@ func getFilenameFromURL(url string) (string, error) {
 
 	defer resp.Body.Close()
 
-	_, params, err := mime.ParseMediaType(resp.Header["Content-Disposition"][0])
+	contentDisposition := resp.Header.Get("Content-Disposition")
+	if contentDisposition == "" {
+		return "", fmt.Errorf("no Content-Disposition header in response from %s", url)
+	}
+
+	_, params, err := mime.ParseMediaType(contentDisposition)
 	if err != nil {
 		return "", err
 	}

management/server/group.go

@@ -6,11 +6,14 @@ import (
 	"fmt"
 	"slices"
 
+	agentNetworkTypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/rs/xid"
 	log "github.com/sirupsen/logrus"
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	routerTypes "github.com/netbirdio/netbird/management/server/networks/routers/types"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
@@ -77,7 +80,8 @@ func (am *DefaultAccountManager) GetGroupByName(ctx context.Context, groupName,
 // CreateGroup object of the peers
 func (am *DefaultAccountManager) CreateGroup(ctx context.Context, accountID, userID string, newGroup *types.Group) error {
 	var eventsToStore []func()
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
+	change := affectedpeers.Change{ChangedGroupIDs: []string{newGroup.ID}}
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		var err error
@@ -90,10 +94,7 @@ func (am *DefaultAccountManager) CreateGroup(ctx context.Context, accountID, use
 		events := am.prepareGroupEvents(ctx, transaction, accountID, userID, newGroup)
 		eventsToStore = append(eventsToStore, events...)
 
-		updateAccountPeers, err = areGroupChangesAffectPeers(ctx, transaction, accountID, []string{newGroup.ID})
-		if err != nil {
-			return err
-		}
+		newGroup.PublicID = xid.New().String()
 
 		if err := transaction.CreateGroup(ctx, newGroup); err != nil {
 			return status.Errorf(status.Internal, "failed to create group: %v", err)
@@ -105,6 +106,11 @@ func (am *DefaultAccountManager) CreateGroup(ctx context.Context, accountID, use
 			}
 		}
 
+		snap, err = affectedpeers.Load(ctx, transaction, accountID, change)
+		if err != nil {
+			return err
+		}
+
 		return transaction.IncrementNetworkSerial(ctx, accountID)
 	})
 	if err != nil {
@@ -115,9 +121,7 @@ func (am *DefaultAccountManager) CreateGroup(ctx context.Context, accountID, use
 		storeEvent()
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -125,7 +129,8 @@ func (am *DefaultAccountManager) CreateGroup(ctx context.Context, accountID, use
 // UpdateGroup object of the peers
 func (am *DefaultAccountManager) UpdateGroup(ctx context.Context, accountID, userID string, newGroup *types.Group) error {
 	var eventsToStore []func()
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
+	change := affectedpeers.Change{ChangedGroupIDs: []string{newGroup.ID}}
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		if err := validateNewGroup(ctx, transaction, accountID, newGroup); err != nil {
@@ -144,27 +149,31 @@ func (am *DefaultAccountManager) UpdateGroup(ctx context.Context, accountID, use
 
 		peersToAdd := util.Difference(newGroup.Peers, oldGroup.Peers)
 		peersToRemove := util.Difference(oldGroup.Peers, newGroup.Peers)
-
-		for _, peerID := range peersToAdd {
-			if err := transaction.AddPeerToGroup(ctx, accountID, peerID, newGroup.ID); err != nil {
-				return status.Errorf(status.Internal, "failed to add peer %s to group %s: %v", peerID, newGroup.ID, err)
-			}
-		}
-		for _, peerID := range peersToRemove {
-			if err := transaction.RemovePeerFromGroup(ctx, peerID, newGroup.ID); err != nil {
-				return status.Errorf(status.Internal, "failed to remove peer %s from group %s: %v", peerID, newGroup.ID, err)
-			}
-		}
-
-		updateAccountPeers, err = areGroupChangesAffectPeers(ctx, transaction, accountID, []string{newGroup.ID})
-		if err != nil {
+		if err = syncGroupMembership(ctx, transaction, accountID, newGroup.ID, peersToAdd, peersToRemove); err != nil {
 			return err
 		}
 
+		newGroup.PublicID = oldGroup.PublicID
+
 		if err = transaction.UpdateGroup(ctx, newGroup); err != nil {
 			return err
 		}
 
+		if err = am.reconcileIPv6ForGroupChanges(ctx, transaction, accountID, []string{newGroup.ID}); err != nil {
+			return err
+		}
+
+		// A membership change does not alter which entities reference the group, so
+		// the dependency walk runs once against the post-change snapshot. The new
+		// members are already in the snapshot's index; the removed members are
+		// carried separately and folded in only when the group is linked.
+		if len(peersToRemove) > 0 {
+			change.RemovedPeersByGroup = map[string][]string{newGroup.ID: peersToRemove}
+		}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
+			return err
+		}
+
 		return transaction.IncrementNetworkSerial(ctx, accountID)
 	})
 	if err != nil {
@@ -175,30 +184,48 @@ func (am *DefaultAccountManager) UpdateGroup(ctx context.Context, accountID, use
 		storeEvent()
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
 
+// syncGroupMembership applies the peer membership delta for a group within a transaction.
+func syncGroupMembership(ctx context.Context, transaction store.Store, accountID, groupID string, peersToAdd, peersToRemove []string) error {
+	for _, peerID := range peersToAdd {
+		if err := transaction.AddPeerToGroup(ctx, accountID, peerID, groupID); err != nil {
+			return status.Errorf(status.Internal, "failed to add peer %s to group %s: %v", peerID, groupID, err)
+		}
+	}
+	for _, peerID := range peersToRemove {
+		if err := transaction.RemovePeerFromGroup(ctx, peerID, groupID); err != nil {
+			return status.Errorf(status.Internal, "failed to remove peer %s from group %s: %v", peerID, groupID, err)
+		}
+	}
+	return nil
+}
+
 // CreateGroups adds new groups to the account.
 // Note: This function does not acquire the global lock.
 // It is the caller's responsibility to ensure proper locking is in place before invoking this method.
 // This method will not create group peer membership relations. Use AddPeerToGroup or RemovePeerFromGroup methods for that.
 func (am *DefaultAccountManager) CreateGroups(ctx context.Context, accountID, userID string, groups []*types.Group) error {
 	var eventsToStore []func()
-	var updateAccountPeers bool
+	var snaps []*affectedpeers.Snapshot
+	var changes []affectedpeers.Change
 
+	var err error
 	var globalErr error
-	groupIDs := make([]string, 0, len(groups))
+	createdCount := 0
 	for _, newGroup := range groups {
-		err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-			if err := validateNewGroup(ctx, transaction, accountID, newGroup); err != nil {
+		change := affectedpeers.Change{ChangedGroupIDs: []string{newGroup.ID}}
+		var snap *affectedpeers.Snapshot
+		err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
+			if err = validateNewGroup(ctx, transaction, accountID, newGroup); err != nil {
 				return err
 			}
 
 			newGroup.AccountID = accountID
+			newGroup.PublicID = xid.New().String()
 
 			if err := transaction.CreateGroup(ctx, newGroup); err != nil {
 				return err
@@ -208,36 +235,31 @@ func (am *DefaultAccountManager) CreateGroups(ctx context.Context, accountID, us
 				return err
 			}
 
-			groupIDs = append(groupIDs, newGroup.ID)
-
 			events := am.prepareGroupEvents(ctx, transaction, accountID, userID, newGroup)
 			eventsToStore = append(eventsToStore, events...)
 
-			return nil
+			snap, err = affectedpeers.Load(ctx, transaction, accountID, change)
+			return err
 		})
 		if err != nil {
 			log.WithContext(ctx).Errorf("failed to update group %s: %v", newGroup.ID, err)
-			if len(groupIDs) == 1 {
+			if createdCount == 0 {
 				return err
 			}
 			globalErr = errors.Join(globalErr, err)
 			// continue updating other groups
+			continue
 		}
-	}
-
-	var err error
-	updateAccountPeers, err = areGroupChangesAffectPeers(ctx, am.Store, accountID, groupIDs)
-	if err != nil {
-		return err
+		createdCount++
+		snaps = append(snaps, snap)
+		changes = append(changes, change)
 	}
 
 	for _, storeEvent := range eventsToStore {
 		storeEvent()
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	go am.dispatchAffected(ctx, accountID, snaps, changes)
 
 	return globalErr
 }
@@ -248,62 +270,71 @@ func (am *DefaultAccountManager) CreateGroups(ctx context.Context, accountID, us
 // This method will not create group peer membership relations. Use AddPeerToGroup or RemovePeerFromGroup methods for that.
 func (am *DefaultAccountManager) UpdateGroups(ctx context.Context, accountID, userID string, groups []*types.Group) error {
 	var eventsToStore []func()
-	var updateAccountPeers bool
+	var snaps []*affectedpeers.Snapshot
+	var changes []affectedpeers.Change
 
 	var globalErr error
-	groupIDs := make([]string, 0, len(groups))
 	for _, newGroup := range groups {
-		err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-			var err error
-			if err = validateNewGroup(ctx, transaction, accountID, newGroup); err != nil {
-				return err
-			}
-
-			newGroup.AccountID = accountID
-
-			if err = transaction.UpdateGroup(ctx, newGroup); err != nil {
-				return err
-			}
-
-			err = transaction.IncrementNetworkSerial(ctx, accountID)
-			if err != nil {
-				return err
-			}
-
-			events := am.prepareGroupEvents(ctx, transaction, accountID, userID, newGroup)
-			eventsToStore = append(eventsToStore, events...)
-
-			groupIDs = append(groupIDs, newGroup.ID)
-
-			return nil
-		})
+		change := affectedpeers.Change{ChangedGroupIDs: []string{newGroup.ID}}
+		events, snap, err := am.updateSingleGroup(ctx, accountID, userID, newGroup, change)
 		if err != nil {
 			log.WithContext(ctx).Errorf("failed to update group %s: %v", newGroup.ID, err)
 			if len(groups) == 1 {
 				return err
 			}
 			globalErr = errors.Join(globalErr, err)
-			// continue updating other groups
+			continue
 		}
-	}
-
-	var err error
-	updateAccountPeers, err = areGroupChangesAffectPeers(ctx, am.Store, accountID, groupIDs)
-	if err != nil {
-		return err
+		eventsToStore = append(eventsToStore, events...)
+		snaps = append(snaps, snap)
+		changes = append(changes, change)
 	}
 
 	for _, storeEvent := range eventsToStore {
 		storeEvent()
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	go am.dispatchAffected(ctx, accountID, snaps, changes)
 
 	return globalErr
 }
 
+func (am *DefaultAccountManager) updateSingleGroup(ctx context.Context, accountID, userID string, newGroup *types.Group, change affectedpeers.Change) ([]func(), *affectedpeers.Snapshot, error) {
+	var events []func()
+	var snap *affectedpeers.Snapshot
+	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
+		if err := validateNewGroup(ctx, transaction, accountID, newGroup); err != nil {
+			return err
+		}
+
+		newGroup.AccountID = accountID
+
+		oldGroup, err := transaction.GetGroupByID(ctx, store.LockingStrengthNone, accountID, newGroup.ID)
+		if err != nil {
+			return err
+		}
+		newGroup.PublicID = oldGroup.PublicID
+
+		if err := transaction.UpdateGroup(ctx, newGroup); err != nil {
+			return err
+		}
+
+		if err := am.reconcileIPv6ForGroupChanges(ctx, transaction, accountID, []string{newGroup.ID}); err != nil {
+			return err
+		}
+
+		if err := transaction.IncrementNetworkSerial(ctx, accountID); err != nil {
+			return err
+		}
+
+		events = am.prepareGroupEvents(ctx, transaction, accountID, userID, newGroup)
+
+		snap, err = affectedpeers.Load(ctx, transaction, accountID, change)
+		return err
+	})
+	return events, snap, err
+}
+
 // prepareGroupEvents prepares a list of event functions to be stored.
 func (am *DefaultAccountManager) prepareGroupEvents(ctx context.Context, transaction store.Store, accountID, userID string, newGroup *types.Group) []func() {
 	var eventsToStore []func()
@@ -396,6 +427,8 @@ func (am *DefaultAccountManager) DeleteGroups(ctx context.Context, accountID, us
 	var allErrors error
 	var groupIDsToDelete []string
 	var deletedGroups []*types.Group
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 
 	extraSettings, err := am.settingsManager.GetExtraSettings(ctx, accountID)
 	if err != nil {
@@ -403,30 +436,31 @@ func (am *DefaultAccountManager) DeleteGroups(ctx context.Context, accountID, us
 	}
 
 	err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		for _, groupID := range groupIDs {
-			group, err := transaction.GetGroupByID(ctx, store.LockingStrengthNone, accountID, groupID)
-			if err != nil {
-				allErrors = errors.Join(allErrors, err)
-				continue
-			}
-
-			if err = validateDeleteGroup(ctx, transaction, group, userID, extraSettings.FlowGroups); err != nil {
-				allErrors = errors.Join(allErrors, err)
-				continue
-			}
-
-			groupIDsToDelete = append(groupIDsToDelete, groupID)
-			deletedGroups = append(deletedGroups, group)
+		deletedGroups, allErrors = collectDeletableGroups(ctx, transaction, accountID, userID, groupIDs, extraSettings.FlowGroups)
+		for _, group := range deletedGroups {
+			groupIDsToDelete = append(groupIDsToDelete, group.ID)
 		}
 
 		if len(groupIDsToDelete) == 0 {
 			return allErrors
 		}
 
+		// Delete: compute affected peers from the PRE-delete state. The groups,
+		// their members and the entities referencing them still exist, so a plain
+		// Load+Expand captures everyone — no removed-peer folding needed.
+		change = affectedpeers.Change{ChangedGroupIDs: groupIDsToDelete}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
+			return err
+		}
+
 		if err = transaction.DeleteGroups(ctx, accountID, groupIDsToDelete); err != nil {
 			return err
 		}
 
+		if err = am.reconcileIPv6ForGroupChanges(ctx, transaction, accountID, groupIDsToDelete); err != nil {
+			return err
+		}
+
 		return transaction.IncrementNetworkSerial(ctx, accountID)
 	})
 	if err != nil {
@@ -437,21 +471,52 @@ func (am *DefaultAccountManager) DeleteGroups(ctx context.Context, accountID, us
 		am.StoreEvent(ctx, userID, group.ID, accountID, activity.GroupDeleted, group.EventMeta())
 	}
 
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
+
 	return allErrors
 }
 
+// collectDeletableGroups loads and validates each group for deletion, returning
+// the groups that may be deleted and the joined validation errors for the rest.
+func collectDeletableGroups(ctx context.Context, transaction store.Store, accountID, userID string, groupIDs, flowGroups []string) ([]*types.Group, error) {
+	var deletable []*types.Group
+	var allErrors error
+	for _, groupID := range groupIDs {
+		group, err := transaction.GetGroupByID(ctx, store.LockingStrengthNone, accountID, groupID)
+		if err != nil {
+			allErrors = errors.Join(allErrors, err)
+			continue
+		}
+		if err = validateDeleteGroup(ctx, transaction, group, userID, flowGroups); err != nil {
+			allErrors = errors.Join(allErrors, err)
+			continue
+		}
+		deletable = append(deletable, group)
+	}
+	return deletable, allErrors
+}
+
 // GroupAddPeer appends peer to the group
 func (am *DefaultAccountManager) GroupAddPeer(ctx context.Context, accountID, groupID, peerID string) error {
-	var updateAccountPeers bool
-	var err error
+	var snap *affectedpeers.Snapshot
+	// A membership change affects only the peer itself and the opposite side of THIS
+	// group's policies — not the group's other members, and not the peer's other
+	// groups. LinkGroups walks only this group (matched, not expanded); OutputPeerIDs
+	// refreshes the peer without seeding its other group memberships. For an
+	// intra-group policy the opposite side is the group, so its members still refresh.
+	change := affectedpeers.Change{OutputPeerIDs: []string{peerID}, LinkGroups: []string{groupID}}
 
-	err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		updateAccountPeers, err = areGroupChangesAffectPeers(ctx, transaction, accountID, []string{groupID})
-		if err != nil {
+	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
+		if err := transaction.AddPeerToGroup(ctx, accountID, peerID, groupID); err != nil {
 			return err
 		}
 
-		if err = transaction.AddPeerToGroup(ctx, accountID, peerID, groupID); err != nil {
+		if err := am.reconcileIPv6ForGroupChanges(ctx, transaction, accountID, []string{groupID}); err != nil {
+			return err
+		}
+
+		var err error
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -461,9 +526,7 @@ func (am *DefaultAccountManager) GroupAddPeer(ctx context.Context, accountID, gr
 		return err
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -471,8 +534,9 @@ func (am *DefaultAccountManager) GroupAddPeer(ctx context.Context, accountID, gr
 // GroupAddResource appends resource to the group
 func (am *DefaultAccountManager) GroupAddResource(ctx context.Context, accountID, groupID string, resource types.Resource) error {
 	var group *types.Group
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
 	var err error
+	change := affectedpeers.Change{ChangedGroupIDs: []string{groupID}}
 
 	err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		group, err = transaction.GetGroupByID(context.Background(), store.LockingStrengthUpdate, accountID, groupID)
@@ -484,12 +548,11 @@ func (am *DefaultAccountManager) GroupAddResource(ctx context.Context, accountID
 			return nil
 		}
 
-		updateAccountPeers, err = areGroupChangesAffectPeers(ctx, transaction, accountID, []string{groupID})
-		if err != nil {
+		if err = transaction.UpdateGroup(ctx, group); err != nil {
 			return err
 		}
 
-		if err = transaction.UpdateGroup(ctx, group); err != nil {
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -499,25 +562,31 @@ func (am *DefaultAccountManager) GroupAddResource(ctx context.Context, accountID
 		return err
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
 
 // GroupDeletePeer removes peer from the group
 func (am *DefaultAccountManager) GroupDeletePeer(ctx context.Context, accountID, groupID, peerID string) error {
-	var updateAccountPeers bool
-	var err error
+	var snap *affectedpeers.Snapshot
+	// Same as GroupAddPeer: the removed peer and the opposite side of THIS group's
+	// policies refresh, not the group's other members or the peer's other groups. The
+	// peer is no longer in the group's index, but LinkGroups still drives the
+	// opposite-side walk, and OutputPeerIDs refreshes the removed peer itself.
+	change := affectedpeers.Change{OutputPeerIDs: []string{peerID}, LinkGroups: []string{groupID}}
 
-	err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		updateAccountPeers, err = areGroupChangesAffectPeers(ctx, transaction, accountID, []string{groupID})
-		if err != nil {
+	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
+		if err := transaction.RemovePeerFromGroup(ctx, peerID, groupID); err != nil {
 			return err
 		}
 
-		if err = transaction.RemovePeerFromGroup(ctx, peerID, groupID); err != nil {
+		if err := am.reconcileIPv6ForGroupChanges(ctx, transaction, accountID, []string{groupID}); err != nil {
+			return err
+		}
+
+		var err error
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -527,9 +596,7 @@ func (am *DefaultAccountManager) GroupDeletePeer(ctx context.Context, accountID,
 		return err
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -537,8 +604,9 @@ func (am *DefaultAccountManager) GroupDeletePeer(ctx context.Context, accountID,
 // GroupDeleteResource removes resource from the group
 func (am *DefaultAccountManager) GroupDeleteResource(ctx context.Context, accountID, groupID string, resource types.Resource) error {
 	var group *types.Group
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
 	var err error
+	change := affectedpeers.Change{ChangedGroupIDs: []string{groupID}}
 
 	err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		group, err = transaction.GetGroupByID(context.Background(), store.LockingStrengthUpdate, accountID, groupID)
@@ -550,8 +618,9 @@ func (am *DefaultAccountManager) GroupDeleteResource(ctx context.Context, accoun
 			return nil
 		}
 
-		updateAccountPeers, err = areGroupChangesAffectPeers(ctx, transaction, accountID, []string{groupID})
-		if err != nil {
+		// Load before persisting the removal, so the snapshot still maps the group
+		// to the resource and the bridge can reach its routing peers.
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -565,9 +634,7 @@ func (am *DefaultAccountManager) GroupDeleteResource(ctx context.Context, accoun
 		return err
 	}
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -646,6 +713,14 @@ func validateDeleteGroup(ctx context.Context, transaction store.Store, group *ty
 		return &GroupLinkError{"network router", linkedRouter.ID}
 	}
 
+	if isLinked, linkedService := isGroupLinkedToReverseProxyService(ctx, transaction, group.AccountID, group.ID); isLinked {
+		return &GroupLinkError{"reverse proxy service", linkedService.Domain}
+	}
+
+	if isLinked, linkedPolicy := isGroupLinkedToAgentNetworkPolicy(ctx, transaction, group.AccountID, group.ID); isLinked {
+		return &GroupLinkError{"agent network policy", linkedPolicy.Name}
+	}
+
 	return checkGroupLinkedToSettings(ctx, transaction, group)
 }
 
@@ -777,50 +852,144 @@ func isGroupLinkedToNetworkRouter(ctx context.Context, transaction store.Store,
 	return false, nil
 }
 
+// isGroupLinkedToReverseProxyService checks if a group is used as an access group
+// of a private reverse proxy service or as a bearer-auth distribution group.
+func isGroupLinkedToReverseProxyService(ctx context.Context, transaction store.Store, accountID string, groupID string) (bool, *service.Service) {
+	services, err := transaction.GetAccountServices(ctx, store.LockingStrengthNone, accountID)
+	if err != nil {
+		log.WithContext(ctx).Errorf("error retrieving reverse proxy services while checking group linkage: %v", err)
+		return false, nil
+	}
+
+	for _, svc := range services {
+		if svc.Private && slices.Contains(svc.AccessGroups, groupID) {
+			return true, svc
+		}
+		if svc.Auth.BearerAuth != nil && svc.Auth.BearerAuth.Enabled && slices.Contains(svc.Auth.BearerAuth.DistributionGroups, groupID) {
+			return true, svc
+		}
+	}
+	return false, nil
+}
+
+// isGroupLinkedToAgentNetworkPolicy checks if a group is used as a source group by any
+// agent network policy in the account.
+func isGroupLinkedToAgentNetworkPolicy(ctx context.Context, transaction store.Store, accountID string, groupID string) (bool, *agentNetworkTypes.Policy) {
+	policies, err := transaction.GetAccountAgentNetworkPolicies(ctx, store.LockingStrengthNone, accountID)
+	if err != nil {
+		log.WithContext(ctx).Errorf("error retrieving agent network policies while checking group linkage: %v", err)
+		return false, nil
+	}
+
+	for _, policy := range policies {
+		if policy == nil {
+			continue
+		}
+		if slices.Contains(policy.SourceGroups, groupID) {
+			return true, policy
+		}
+	}
+	return false, nil
+}
+
 // areGroupChangesAffectPeers checks if any changes to the specified groups will affect peers.
+// It fetches each collection once and checks all groupIDs against them in memory.
 func areGroupChangesAffectPeers(ctx context.Context, transaction store.Store, accountID string, groupIDs []string) (bool, error) {
 	if len(groupIDs) == 0 {
 		return false, nil
 	}
 
+	groupSet := make(map[string]struct{}, len(groupIDs))
+	for _, id := range groupIDs {
+		groupSet[id] = struct{}{}
+	}
+
+	if affected, err := dnsSettingsReferenceGroups(ctx, transaction, accountID, groupSet); affected || err != nil {
+		return affected, err
+	}
+	if affected, err := nameServersReferenceGroups(ctx, transaction, accountID, groupSet); affected || err != nil {
+		return affected, err
+	}
+	if affected, err := policiesReferenceGroups(ctx, transaction, accountID, groupSet); affected || err != nil {
+		return affected, err
+	}
+	if affected, err := routesReferenceGroups(ctx, transaction, accountID, groupSet); affected || err != nil {
+		return affected, err
+	}
+	if affected, err := networkRoutersReferenceGroups(ctx, transaction, accountID, groupSet); affected || err != nil {
+		return affected, err
+	}
+
+	return false, nil
+}
+
+func dnsSettingsReferenceGroups(ctx context.Context, transaction store.Store, accountID string, groupSet map[string]struct{}) (bool, error) {
 	dnsSettings, err := transaction.GetAccountDNSSettings(ctx, store.LockingStrengthNone, accountID)
 	if err != nil {
 		return false, err
 	}
-
-	for _, groupID := range groupIDs {
-		if slices.Contains(dnsSettings.DisabledManagementGroups, groupID) {
-			return true, nil
-		}
-		if linked, _ := isGroupLinkedToDns(ctx, transaction, accountID, groupID); linked {
-			return true, nil
-		}
-		if linked, _ := isGroupLinkedToPolicy(ctx, transaction, accountID, groupID); linked {
-			return true, nil
-		}
-		if linked, _ := isGroupLinkedToRoute(ctx, transaction, accountID, groupID); linked {
-			return true, nil
-		}
-		if linked, _ := isGroupLinkedToNetworkRouter(ctx, transaction, accountID, groupID); linked {
-			return true, nil
-		}
-	}
-
-	return false, nil
+	return anyInSet(dnsSettings.DisabledManagementGroups, groupSet), nil
 }
 
-// anyGroupHasPeersOrResources checks if any of the given groups in the account have peers or resources.
-func anyGroupHasPeersOrResources(ctx context.Context, transaction store.Store, accountID string, groupIDs []string) (bool, error) {
-	groups, err := transaction.GetGroupsByIDs(ctx, store.LockingStrengthNone, accountID, groupIDs)
+func nameServersReferenceGroups(ctx context.Context, transaction store.Store, accountID string, groupSet map[string]struct{}) (bool, error) {
+	nameServerGroups, err := transaction.GetAccountNameServerGroups(ctx, store.LockingStrengthNone, accountID)
 	if err != nil {
 		return false, err
 	}
-
-	for _, group := range groups {
-		if group.HasPeers() || group.HasResources() {
+	for _, ns := range nameServerGroups {
+		if anyInSet(ns.Groups, groupSet) {
 			return true, nil
 		}
 	}
-
 	return false, nil
 }
+
+func policiesReferenceGroups(ctx context.Context, transaction store.Store, accountID string, groupSet map[string]struct{}) (bool, error) {
+	policies, err := transaction.GetAccountPolicies(ctx, store.LockingStrengthNone, accountID)
+	if err != nil {
+		return false, err
+	}
+	for _, policy := range policies {
+		for _, rule := range policy.Rules {
+			if anyInSet(rule.Sources, groupSet) || anyInSet(rule.Destinations, groupSet) {
+				return true, nil
+			}
+		}
+	}
+	return false, nil
+}
+
+func routesReferenceGroups(ctx context.Context, transaction store.Store, accountID string, groupSet map[string]struct{}) (bool, error) {
+	routes, err := transaction.GetAccountRoutes(ctx, store.LockingStrengthNone, accountID)
+	if err != nil {
+		return false, err
+	}
+	for _, r := range routes {
+		if anyInSet(r.Groups, groupSet) || anyInSet(r.PeerGroups, groupSet) || anyInSet(r.AccessControlGroups, groupSet) {
+			return true, nil
+		}
+	}
+	return false, nil
+}
+
+func networkRoutersReferenceGroups(ctx context.Context, transaction store.Store, accountID string, groupSet map[string]struct{}) (bool, error) {
+	routers, err := transaction.GetNetworkRoutersByAccountID(ctx, store.LockingStrengthNone, accountID)
+	if err != nil {
+		return false, err
+	}
+	for _, router := range routers {
+		if anyInSet(router.PeerGroups, groupSet) {
+			return true, nil
+		}
+	}
+	return false, nil
+}
+
+func anyInSet(ids []string, set map[string]struct{}) bool {
+	for _, id := range ids {
+		if _, ok := set[id]; ok {
+			return true
+		}
+	}
+	return false
+}

management/server/group_ipv6_test.go

@@ -0,0 +1,124 @@
+package server
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// TestGroupIPv6Assignment verifies that peers gain or lose IPv6 addresses
+// when they are added to or removed from an IPv6-enabled group.
+func TestGroupIPv6Assignment(t *testing.T) {
+	am, _, err := createManager(t)
+	require.NoError(t, err)
+
+	ctx := context.Background()
+	userID := groupAdminUserID
+
+	account, err := createAccount(am, "ipv6-grp-test", userID, "ipv6test.example.com")
+	require.NoError(t, err)
+
+	// Allocate IPv6 subnet for the account
+	account.Network.NetV6 = types.AllocateIPv6Subnet()
+	require.NoError(t, am.Store.SaveAccount(ctx, account))
+
+	// Create setup key
+	setupKey, err := am.CreateSetupKey(ctx, account.Id, "ipv6-key", types.SetupKeyReusable, time.Hour, nil, 999, userID, false, false)
+	require.NoError(t, err)
+
+	// Create an IPv6-enabled group
+	ipv6GroupID := "ipv6-enabled-grp"
+	err = am.CreateGroup(ctx, account.Id, userID, &types.Group{
+		ID:     ipv6GroupID,
+		Name:   "IPv6 Enabled",
+		Issued: types.GroupIssuedAPI,
+		Peers:  []string{},
+	})
+	require.NoError(t, err)
+
+	// Enable IPv6 on that group
+	settings, err := am.Store.GetAccountSettings(ctx, store.LockingStrengthNone, account.Id)
+	require.NoError(t, err)
+	settings.IPv6EnabledGroups = []string{ipv6GroupID}
+	require.NoError(t, am.Store.SaveAccountSettings(ctx, account.Id, settings))
+
+	// Register a peer (will be in "All" group, not the IPv6 group)
+	key, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err)
+
+	peer, _, _, _, err := am.AddPeer(ctx, "", setupKey.Key, "", &nbpeer.Peer{
+		Key:  key.PublicKey().String(),
+		Meta: nbpeer.PeerSystemMeta{Hostname: "ipv6-test-host"},
+	}, false)
+	require.NoError(t, err)
+	assert.False(t, peer.IPv6.IsValid(), "peer should not have IPv6 before joining an IPv6-enabled group")
+
+	t.Run("GroupAddPeer assigns IPv6", func(t *testing.T) {
+		err := am.GroupAddPeer(ctx, account.Id, ipv6GroupID, peer.ID)
+		require.NoError(t, err)
+
+		p, err := am.Store.GetPeerByID(ctx, store.LockingStrengthNone, account.Id, peer.ID)
+		require.NoError(t, err)
+		assert.True(t, p.IPv6.IsValid(), "peer should have an IPv6 address after joining the group")
+	})
+
+	t.Run("GroupDeletePeer clears IPv6", func(t *testing.T) {
+		err := am.GroupDeletePeer(ctx, account.Id, ipv6GroupID, peer.ID)
+		require.NoError(t, err)
+
+		p, err := am.Store.GetPeerByID(ctx, store.LockingStrengthNone, account.Id, peer.ID)
+		require.NoError(t, err)
+		assert.False(t, p.IPv6.IsValid(), "peer should not have IPv6 after removal from the group")
+	})
+
+	t.Run("UpdateGroup with peer addition assigns IPv6", func(t *testing.T) {
+		grp, err := am.Store.GetGroupByID(ctx, store.LockingStrengthNone, account.Id, ipv6GroupID)
+		require.NoError(t, err)
+
+		grp.Peers = append(grp.Peers, peer.ID)
+		err = am.UpdateGroup(ctx, account.Id, userID, grp)
+		require.NoError(t, err)
+
+		p, err := am.Store.GetPeerByID(ctx, store.LockingStrengthNone, account.Id, peer.ID)
+		require.NoError(t, err)
+		assert.True(t, p.IPv6.IsValid(), "peer should have IPv6 after UpdateGroup adds it")
+	})
+
+	t.Run("UpdateGroup with peer removal clears IPv6", func(t *testing.T) {
+		grp, err := am.Store.GetGroupByID(ctx, store.LockingStrengthNone, account.Id, ipv6GroupID)
+		require.NoError(t, err)
+
+		grp.Peers = []string{}
+		err = am.UpdateGroup(ctx, account.Id, userID, grp)
+		require.NoError(t, err)
+
+		p, err := am.Store.GetPeerByID(ctx, store.LockingStrengthNone, account.Id, peer.ID)
+		require.NoError(t, err)
+		assert.False(t, p.IPv6.IsValid(), "peer should lose IPv6 after UpdateGroup removes it")
+	})
+
+	t.Run("non-IPv6 group changes do not affect IPv6", func(t *testing.T) {
+		err := am.CreateGroup(ctx, account.Id, userID, &types.Group{
+			ID:     "regular-grp",
+			Name:   "Regular Group",
+			Issued: types.GroupIssuedAPI,
+			Peers:  []string{},
+		})
+		require.NoError(t, err)
+
+		err = am.GroupAddPeer(ctx, account.Id, "regular-grp", peer.ID)
+		require.NoError(t, err)
+
+		p, err := am.Store.GetPeerByID(ctx, store.LockingStrengthNone, account.Id, peer.ID)
+		require.NoError(t, err)
+		assert.False(t, p.IPv6.IsValid(), "peer should not get IPv6 from a non-IPv6 group")
+	})
+}

management/server/group_test.go

@@ -5,20 +5,21 @@ import (
 	"encoding/binary"
 	"errors"
 	"fmt"
-	"net"
 	"net/netip"
 	"strconv"
 	"sync"
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/google/uuid"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"golang.org/x/exp/maps"
 
 	nbdns "github.com/netbirdio/netbird/dns"
+	agentNetworkTypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/netbirdio/netbird/management/server/groups"
 	"github.com/netbirdio/netbird/management/server/networks"
 	"github.com/netbirdio/netbird/management/server/networks/resources"
@@ -125,6 +126,21 @@ func TestDefaultAccountManager_DeleteGroup(t *testing.T) {
 			"grp-for-integration",
 			"only service users with admin power can delete integration group",
 		},
+		{
+			"agent network policy",
+			"grp-for-agent-network-policy",
+			"agent network policy",
+		},
+		{
+			"reverse proxy private service access group",
+			"grp-for-rp-private",
+			"reverse proxy service",
+		},
+		{
+			"reverse proxy bearer distribution group",
+			"grp-for-rp-bearer",
+			"reverse proxy service",
+		},
 	}
 
 	for _, testCase := range testCases {
@@ -218,6 +234,17 @@ func TestDefaultAccountManager_DeleteGroups(t *testing.T) {
 			groupIDs:        []string{"grp-for-integration"},
 			expectedReasons: []string{"only service users with admin power can delete integration group"},
 		},
+		{
+			name:            "agent network policy",
+			groupIDs:        []string{"grp-for-agent-network-policy"},
+			expectedReasons: []string{"agent network policy"},
+		},
+		{
+			name:               "reverse proxy services",
+			groupIDs:           []string{"grp-for-rp-private", "grp-for-rp-bearer"},
+			expectedReasons:    []string{"reverse proxy service", "reverse proxy service"},
+			expectedNotDeleted: []string{"grp-for-rp-private", "grp-for-rp-bearer"},
+		},
 		{
 			name:            "successfully delete multiple groups",
 			groupIDs:        []string{"group-1", "group-2"},
@@ -285,6 +312,65 @@ func TestDefaultAccountManager_DeleteGroups(t *testing.T) {
 	}
 }
 
+func TestDefaultAccountManager_DeleteGroupUnlinkedFromReverseProxyService(t *testing.T) {
+	am, _, err := createManager(t)
+	require.NoError(t, err, "Failed to create account manager")
+
+	_, account, err := initTestGroupAccount(am)
+	require.NoError(t, err, "Failed to init testing account")
+
+	deletableGroups := []*types.Group{
+		{
+			ID:        "grp-rp-bearer-disabled",
+			AccountID: account.Id,
+			Name:      "Group only in a disabled bearer auth",
+			Issued:    types.GroupIssuedAPI,
+			Peers:     make([]string, 0),
+		},
+		{
+			ID:        "grp-rp-nonprivate-access",
+			AccountID: account.Id,
+			Name:      "Group only in a non-private service's access groups",
+			Issued:    types.GroupIssuedAPI,
+			Peers:     make([]string, 0),
+		},
+	}
+	for _, group := range deletableGroups {
+		require.NoError(t, am.CreateGroup(context.Background(), account.Id, groupAdminUserID, group))
+	}
+
+	// Disabled bearer auth and stale access groups on a non-private service
+	// are inert configuration and must not block group deletion.
+	services := []*rpservice.Service{
+		{
+			ID:        "rp-svc-bearer-disabled",
+			AccountID: account.Id,
+			Domain:    "bearer-disabled.services.example.com",
+			Auth: rpservice.AuthConfig{
+				BearerAuth: &rpservice.BearerAuthConfig{
+					Enabled:            false,
+					DistributionGroups: []string{"grp-rp-bearer-disabled"},
+				},
+			},
+		},
+		{
+			ID:           "rp-svc-nonprivate-access",
+			AccountID:    account.Id,
+			Domain:       "nonprivate.services.example.com",
+			Private:      false,
+			AccessGroups: []string{"grp-rp-nonprivate-access"},
+		},
+	}
+	for _, svc := range services {
+		require.NoError(t, am.Store.CreateService(context.Background(), svc))
+	}
+
+	for _, group := range deletableGroups {
+		err = am.DeleteGroup(context.Background(), account.Id, groupAdminUserID, group.ID)
+		assert.NoError(t, err, "group %s is not referenced by an active reverse proxy gate and should be deletable", group.ID)
+	}
+}
+
 func TestDefaultAccountManager_DeleteGroupLinkedToFlowGroup(t *testing.T) {
 	am, _, err := createManager(t)
 	require.NoError(t, err)
@@ -406,6 +492,30 @@ func initTestGroupAccount(am *DefaultAccountManager) (*DefaultAccountManager, *t
 		Peers:     make([]string, 0),
 	}
 
+	groupForAgentNetworkPolicy := &types.Group{
+		ID:        "grp-for-agent-network-policy",
+		AccountID: "account-id",
+		Name:      "Group for agent network policies",
+		Issued:    types.GroupIssuedAPI,
+		Peers:     make([]string, 0),
+	}
+
+	groupForRPPrivate := &types.Group{
+		ID:        "grp-for-rp-private",
+		AccountID: "account-id",
+		Name:      "Group for private reverse proxy service",
+		Issued:    types.GroupIssuedAPI,
+		Peers:     make([]string, 0),
+	}
+
+	groupForRPBearer := &types.Group{
+		ID:        "grp-for-rp-bearer",
+		AccountID: "account-id",
+		Name:      "Group for bearer reverse proxy service",
+		Issued:    types.GroupIssuedAPI,
+		Peers:     make([]string, 0),
+	}
+
 	routeResource := &route.Route{
 		ID:     "example route",
 		Groups: []string{groupForRoute.ID},
@@ -461,6 +571,66 @@ func initTestGroupAccount(am *DefaultAccountManager) (*DefaultAccountManager, *t
 	_ = am.CreateGroup(context.Background(), accountID, groupAdminUserID, groupForSetupKeys)
 	_ = am.CreateGroup(context.Background(), accountID, groupAdminUserID, groupForUsers)
 	_ = am.CreateGroup(context.Background(), accountID, groupAdminUserID, groupForIntegration)
+	_ = am.CreateGroup(context.Background(), accountID, groupAdminUserID, groupForAgentNetworkPolicy)
+	_ = am.CreateGroup(context.Background(), accountID, groupAdminUserID, groupForRPPrivate)
+	_ = am.CreateGroup(context.Background(), accountID, groupAdminUserID, groupForRPBearer)
+
+	agentNetworkPolicy := &agentNetworkTypes.Policy{
+		ID:           "example agent network policy",
+		AccountID:    accountID,
+		Name:         "Example agent network policy",
+		Enabled:      true,
+		SourceGroups: []string{groupForAgentNetworkPolicy.ID},
+	}
+	if err := am.Store.SaveAgentNetworkPolicy(context.Background(), agentNetworkPolicy); err != nil {
+		return nil, nil, err
+	}
+
+	// The decoy services are created first so the linkage check has to scan
+	// past services that do not reference the groups under test.
+	rpServices := []*rpservice.Service{
+		{
+			ID:           "rp-svc-private-decoy",
+			AccountID:    accountID,
+			Domain:       "private-decoy.services.example.com",
+			Private:      true,
+			AccessGroups: []string{"unrelated-group"},
+		},
+		{
+			ID:        "rp-svc-bearer-decoy",
+			AccountID: accountID,
+			Domain:    "bearer-decoy.services.example.com",
+			Auth: rpservice.AuthConfig{
+				BearerAuth: &rpservice.BearerAuthConfig{
+					Enabled:            true,
+					DistributionGroups: []string{"unrelated-group"},
+				},
+			},
+		},
+		{
+			ID:           "rp-svc-private",
+			AccountID:    accountID,
+			Domain:       "private.services.example.com",
+			Private:      true,
+			AccessGroups: []string{groupForRPPrivate.ID},
+		},
+		{
+			ID:        "rp-svc-bearer",
+			AccountID: accountID,
+			Domain:    "bearer.services.example.com",
+			Auth: rpservice.AuthConfig{
+				BearerAuth: &rpservice.BearerAuthConfig{
+					Enabled:            true,
+					DistributionGroups: []string{groupForRPBearer.ID},
+				},
+			},
+		},
+	}
+	for _, svc := range rpServices {
+		if err := am.Store.CreateService(context.Background(), svc); err != nil {
+			return nil, nil, err
+		}
+	}
 
 	acc, err := am.Store.GetAccount(context.Background(), account.Id)
 	if err != nil {
@@ -619,7 +789,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -637,7 +807,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -655,7 +825,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -688,7 +858,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -729,7 +899,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -756,7 +926,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -802,7 +972,7 @@ func TestGroupAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -997,10 +1167,10 @@ func Test_AddPeerAndAddToAll(t *testing.T) {
 	assert.Equal(t, totalPeers, len(account.Peers), "Expected %d peers in account %s, got %d", totalPeers, accountID, len(account.Peers))
 }
 
-func uint32ToIP(n uint32) net.IP {
-	ip := make(net.IP, 4)
-	binary.BigEndian.PutUint32(ip, n)
-	return ip
+func uint32ToIP(n uint32) netip.Addr {
+	var b [4]byte
+	binary.BigEndian.PutUint32(b[:], n)
+	return netip.AddrFrom4(b)
 }
 
 func Test_IncrementNetworkSerial(t *testing.T) {

management/server/groups/manager.go

@@ -88,7 +88,7 @@ func (m *managerImpl) AddResourceToGroupInTransaction(ctx context.Context, trans
 	}
 
 	event := func() {
-		m.accountManager.StoreEvent(ctx, userID, groupID, accountID, activity.ResourceAddedToGroup, group.EventMetaResource(networkResource))
+		m.accountManager.StoreEvent(ctx, userID, groupID, accountID, activity.ResourceAddedToGroup, group.EventMetaResource(types.TwinNetworkResource(networkResource)))
 	}
 
 	return event, nil
@@ -112,7 +112,7 @@ func (m *managerImpl) RemoveResourceFromGroupInTransaction(ctx context.Context,
 	}
 
 	event := func() {
-		m.accountManager.StoreEvent(ctx, userID, groupID, accountID, activity.ResourceRemovedFromGroup, group.EventMetaResource(networkResource))
+		m.accountManager.StoreEvent(ctx, userID, groupID, accountID, activity.ResourceRemovedFromGroup, group.EventMetaResource(types.TwinNetworkResource(networkResource)))
 	}
 
 	return event, nil

management/server/http/handler.go

@@ -16,15 +16,16 @@ import (
 	"github.com/netbirdio/netbird/management/server/types"
 
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/accesslogs"
+	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxytoken"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	reverseproxymanager "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service/manager"
 
 	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
 	idpmanager "github.com/netbirdio/netbird/management/server/idp"
 
-	"github.com/netbirdio/management-integrations/integrations"
-
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
+	agentnetworkhandlers "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/handlers"
 	"github.com/netbirdio/netbird/management/internals/modules/zones"
 	zonesManager "github.com/netbirdio/netbird/management/internals/modules/zones/manager"
 	"github.com/netbirdio/netbird/management/internals/modules/zones/records"
@@ -33,9 +34,7 @@ import (
 	"github.com/netbirdio/netbird/management/server/settings"
 
 	"github.com/netbirdio/netbird/management/server/http/handlers/proxy"
-	"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
 
-	nbpeers "github.com/netbirdio/netbird/management/internals/modules/peers"
 	"github.com/netbirdio/netbird/management/server/auth"
 	"github.com/netbirdio/netbird/management/server/geolocation"
 	nbgroups "github.com/netbirdio/netbird/management/server/groups"
@@ -54,19 +53,14 @@ import (
 	"github.com/netbirdio/netbird/management/server/http/middleware"
 	"github.com/netbirdio/netbird/management/server/http/middleware/bypass"
 	nbinstance "github.com/netbirdio/netbird/management/server/instance"
-	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator"
 	nbnetworks "github.com/netbirdio/netbird/management/server/networks"
 	"github.com/netbirdio/netbird/management/server/networks/resources"
 	"github.com/netbirdio/netbird/management/server/networks/routers"
 	"github.com/netbirdio/netbird/management/server/telemetry"
 )
 
-const (
-	apiPrefix = "/api"
-)
-
 // NewAPIHandler creates the Management service HTTP API handler registering all the available endpoints.
-func NewAPIHandler(ctx context.Context, accountManager account.Manager, networksManager nbnetworks.Manager, resourceManager resources.Manager, routerManager routers.Manager, groupsManager nbgroups.Manager, LocationManager geolocation.Geolocation, authManager auth.Manager, appMetrics telemetry.AppMetrics, integratedValidator integrated_validator.IntegratedValidator, proxyController port_forwarding.Controller, permissionsManager permissions.Manager, peersManager nbpeers.Manager, settingsManager settings.Manager, zManager zones.Manager, rManager records.Manager, networkMapController network_map.Controller, idpManager idpmanager.Manager, serviceManager service.Manager, reverseProxyDomainManager *manager.Manager, reverseProxyAccessLogsManager accesslogs.Manager, proxyGRPCServer *nbgrpc.ProxyServiceServer, trustedHTTPProxies []netip.Prefix, rateLimiter *middleware.APIRateLimiter) (http.Handler, error) {
+func NewAPIHandler(ctx context.Context, router *mux.Router, accountManager account.Manager, networksManager nbnetworks.Manager, resourceManager resources.Manager, routerManager routers.Manager, groupsManager nbgroups.Manager, LocationManager geolocation.Geolocation, authManager auth.Manager, appMetrics telemetry.AppMetrics, permissionsManager permissions.Manager, settingsManager settings.Manager, zManager zones.Manager, rManager records.Manager, networkMapController network_map.Controller, idpManager idpmanager.Manager, serviceManager service.Manager, reverseProxyDomainManager *manager.Manager, reverseProxyAccessLogsManager accesslogs.Manager, proxyGRPCServer *nbgrpc.ProxyServiceServer, trustedHTTPProxies []netip.Prefix, rateLimiter *middleware.APIRateLimiter, isValidChildAccount middleware.IsValidChildAccountFunc, agentNetworkManager agentnetwork.Manager) (http.Handler, error) {
 
 	// Register bypass paths for unauthenticated endpoints
 	if err := bypass.AddBypassPath("/api/instance"); err != nil {
@@ -100,25 +94,16 @@ func NewAPIHandler(ctx context.Context, accountManager account.Manager, networks
 		accountManager.GetUserFromUserAuth,
 		rateLimiter,
 		appMetrics.GetMeter(),
+		isValidChildAccount,
 	)
 
 	corsMiddleware := cors.AllowAll()
 
-	rootRouter := mux.NewRouter()
 	metricsMiddleware := appMetrics.HTTPMiddleware()
 
-	prefix := apiPrefix
-	router := rootRouter.PathPrefix(prefix).Subrouter()
-
 	router.Use(metricsMiddleware.Handler, corsMiddleware.Handler, authMiddleware.Handler)
 
-	if _, err := integrations.RegisterHandlers(ctx, prefix, router, accountManager, integratedValidator, appMetrics.GetMeter(), permissionsManager, peersManager, proxyController, settingsManager); err != nil {
-		return nil, fmt.Errorf("register integrations endpoints: %w", err)
-	}
-
-	// Check if embedded IdP is enabled for instance manager
-	embeddedIdP, embeddedIdpEnabled := idpManager.(*idpmanager.EmbeddedIdPManager)
-	instanceManager, err := nbinstance.NewManager(ctx, accountManager.GetStore(), embeddedIdP)
+	instanceManager, err := nbinstance.NewManager(ctx, accountManager.GetStore(), idpManager)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create instance manager: %w", err)
 	}
@@ -140,21 +125,22 @@ func NewAPIHandler(ctx context.Context, accountManager account.Manager, networks
 	zonesManager.RegisterEndpoints(router, zManager, permissionsManager)
 	recordsManager.RegisterEndpoints(router, rManager, permissionsManager)
 	idp.AddEndpoints(accountManager, router, permissionsManager)
-	instance.AddEndpoints(instanceManager, router)
+	if agentNetworkManager != nil {
+		agentnetworkhandlers.RegisterEndpoints(agentNetworkManager, router)
+	}
+	instance.AddEndpoints(instanceManager, accountManager, router)
 	instance.AddVersionEndpoint(instanceManager, router, permissionsManager)
 	if serviceManager != nil && reverseProxyDomainManager != nil {
 		reverseproxymanager.RegisterEndpoints(serviceManager, *reverseProxyDomainManager, reverseProxyAccessLogsManager, permissionsManager, router)
 	}
+
+	proxytoken.RegisterEndpoints(accountManager.GetStore(), permissionsManager, router)
+
 	// Register OAuth callback handler for proxy authentication
 	if proxyGRPCServer != nil {
 		oauthHandler := proxy.NewAuthCallbackHandler(proxyGRPCServer, trustedHTTPProxies)
 		oauthHandler.RegisterEndpoints(router)
 	}
 
-	// Mount embedded IdP handler at /oauth2 path if configured
-	if embeddedIdpEnabled {
-		rootRouter.PathPrefix("/oauth2").Handler(corsMiddleware.Handler(embeddedIdP.Handler()))
-	}
-
-	return rootRouter, nil
+	return router, nil
 }

management/server/http/handlers/accounts/accounts_handler.go

@@ -4,10 +4,13 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"math"
 	"net/http"
 	"net/netip"
 	"time"
 
+	log "github.com/sirupsen/logrus"
+
 	"github.com/gorilla/mux"
 
 	goversion "github.com/hashicorp/go-version"
@@ -32,7 +35,9 @@ const (
 	// MinNetworkBits is the minimum prefix length for IPv4 network ranges (e.g., /29 gives 8 addresses, /28 gives 16)
 	MinNetworkBitsIPv4 = 28
 	// MinNetworkBitsIPv6 is the minimum prefix length for IPv6 network ranges
-	MinNetworkBitsIPv6      = 120
+	MinNetworkBitsIPv6 = 120
+	// MaxNetworkSizeIPv6 is the largest allowed IPv6 prefix (smallest number)
+	MaxNetworkSizeIPv6      = 48
 	disableAutoUpdate       = "disabled"
 	autoUpdateLatestVersion = "latest"
 )
@@ -79,12 +84,35 @@ func validateMinimumSize(prefix netip.Prefix) error {
 	if addr.Is4() && prefix.Bits() > MinNetworkBitsIPv4 {
 		return status.Errorf(status.InvalidArgument, "network range too small: minimum size is /%d for IPv4", MinNetworkBitsIPv4)
 	}
-	if addr.Is6() && prefix.Bits() > MinNetworkBitsIPv6 {
-		return status.Errorf(status.InvalidArgument, "network range too small: minimum size is /%d for IPv6", MinNetworkBitsIPv6)
+	if addr.Is6() {
+		if prefix.Bits() > MinNetworkBitsIPv6 {
+			return status.Errorf(status.InvalidArgument, "network range too small: minimum size is /%d for IPv6", MinNetworkBitsIPv6)
+		}
+		if prefix.Bits() < MaxNetworkSizeIPv6 {
+			return status.Errorf(status.InvalidArgument, "network range too large: maximum size is /%d for IPv6", MaxNetworkSizeIPv6)
+		}
 	}
 	return nil
 }
 
+func (h *handler) parseAndValidateNetworkRange(ctx context.Context, accountID, userID, rangeStr string, requireV6 bool) (netip.Prefix, error) {
+	prefix, err := netip.ParsePrefix(rangeStr)
+	if err != nil {
+		return netip.Prefix{}, status.Errorf(status.InvalidArgument, "invalid CIDR format: %v", err)
+	}
+	prefix = prefix.Masked()
+	if requireV6 && !prefix.Addr().Is6() {
+		return netip.Prefix{}, status.Errorf(status.InvalidArgument, "network range must be an IPv6 address")
+	}
+	if !requireV6 && prefix.Addr().Is6() {
+		return netip.Prefix{}, status.Errorf(status.InvalidArgument, "network range must be an IPv4 address")
+	}
+	if err := h.validateNetworkRange(ctx, accountID, userID, prefix); err != nil {
+		return netip.Prefix{}, err
+	}
+	return prefix, nil
+}
+
 func (h *handler) validateNetworkRange(ctx context.Context, accountID, userID string, networkRange netip.Prefix) error {
 	if !networkRange.IsValid() {
 		return nil
@@ -120,9 +148,12 @@ func (h *handler) validateCapacity(ctx context.Context, accountID, userID string
 }
 
 func calculateMaxHosts(prefix netip.Prefix) int64 {
-	availableAddresses := prefix.Addr().BitLen() - prefix.Bits()
-	maxHosts := int64(1) << availableAddresses
+	hostBits := prefix.Addr().BitLen() - prefix.Bits()
+	if hostBits >= 63 {
+		return math.MaxInt64
+	}
 
+	maxHosts := int64(1) << hostBits
 	if prefix.Addr().Is4() {
 		maxHosts -= 2 // network and broadcast addresses
 	}
@@ -159,6 +190,24 @@ func (h *handler) getAllAccounts(w http.ResponseWriter, r *http.Request, userAut
 	}
 
 	resp := toAccountResponse(userAuth.AccountId, settings, meta, onboarding)
+
+	// Populate effective network ranges when settings don't have explicit overrides.
+	if resp.Settings.NetworkRange == nil || resp.Settings.NetworkRangeV6 == nil {
+		v4, v6, err := h.settingsManager.GetEffectiveNetworkRanges(r.Context(), userAuth.AccountId)
+		if err != nil {
+			log.WithContext(r.Context()).Warnf("get effective network ranges: %v", err)
+		} else {
+			if resp.Settings.NetworkRange == nil && v4.IsValid() {
+				s := v4.String()
+				resp.Settings.NetworkRange = &s
+			}
+			if resp.Settings.NetworkRangeV6 == nil && v6.IsValid() {
+				s := v6.String()
+				resp.Settings.NetworkRangeV6 = &s
+			}
+		}
+	}
+
 	util.WriteJSONObject(r.Context(), w, []*api.Account{resp})
 }
 
@@ -223,6 +272,30 @@ func (h *handler) updateAccountRequestSettings(req api.PutApiAccountsAccountIdJS
 	if req.Settings.AutoUpdateAlways != nil {
 		returnSettings.AutoUpdateAlways = *req.Settings.AutoUpdateAlways
 	}
+	if req.Settings.LocalMfaEnabled != nil {
+		returnSettings.LocalMfaEnabled = *req.Settings.LocalMfaEnabled
+	}
+	if req.Settings.Ipv6EnabledGroups != nil {
+		returnSettings.IPv6EnabledGroups = *req.Settings.Ipv6EnabledGroups
+	}
+	if req.Settings.MetricsPushEnabled != nil {
+		returnSettings.MetricsPushEnabled = *req.Settings.MetricsPushEnabled
+	}
+	if req.Settings.AgentNetworkOnly != nil {
+		returnSettings.AgentNetworkOnly = *req.Settings.AgentNetworkOnly
+	}
+	if req.Settings.DashboardFeatures != nil {
+		returnSettings.DashboardFeatures = &types.DashboardFeatures{
+			AgentNetwork: req.Settings.DashboardFeatures.AgentNetwork,
+		}
+	}
+
+	if returnSettings.AgentNetworkOnly &&
+		(returnSettings.DashboardFeatures == nil ||
+			returnSettings.DashboardFeatures.AgentNetwork == nil ||
+			!*returnSettings.DashboardFeatures.AgentNetwork) {
+		return nil, status.Errorf(status.InvalidArgument, "agent network only mode requires dashboard_features.agent_network to be enabled")
+	}
 
 	return returnSettings, nil
 }
@@ -248,18 +321,23 @@ func (h *handler) updateAccount(w http.ResponseWriter, r *http.Request, userAuth
 		return
 	}
 	if req.Settings.NetworkRange != nil && *req.Settings.NetworkRange != "" {
-		prefix, err := netip.ParsePrefix(*req.Settings.NetworkRange)
+		prefix, err := h.parseAndValidateNetworkRange(r.Context(), accountID, userAuth.UserId, *req.Settings.NetworkRange, false)
 		if err != nil {
-			util.WriteError(r.Context(), status.Errorf(status.InvalidArgument, "invalid CIDR format: %v", err), w)
-			return
-		}
-		if err := h.validateNetworkRange(r.Context(), accountID, userAuth.UserId, prefix); err != nil {
 			util.WriteError(r.Context(), err, w)
 			return
 		}
 		settings.NetworkRange = prefix
 	}
 
+	if req.Settings.NetworkRangeV6 != nil && *req.Settings.NetworkRangeV6 != "" {
+		prefix, err := h.parseAndValidateNetworkRange(r.Context(), accountID, userAuth.UserId, *req.Settings.NetworkRangeV6, true)
+		if err != nil {
+			util.WriteError(r.Context(), err, w)
+			return
+		}
+		settings.NetworkRangeV6 = prefix
+	}
+
 	var onboarding *types.AccountOnboarding
 	if req.Onboarding != nil {
 		onboarding = &types.AccountOnboarding{
@@ -331,14 +409,27 @@ func toAccountResponse(accountID string, settings *types.Settings, meta *types.A
 		DnsDomain:                       &settings.DNSDomain,
 		AutoUpdateVersion:               &settings.AutoUpdateVersion,
 		AutoUpdateAlways:                &settings.AutoUpdateAlways,
+		Ipv6EnabledGroups:               &settings.IPv6EnabledGroups,
+		MetricsPushEnabled:              &settings.MetricsPushEnabled,
+		AgentNetworkOnly:                &settings.AgentNetworkOnly,
 		EmbeddedIdpEnabled:              &settings.EmbeddedIdpEnabled,
 		LocalAuthDisabled:               &settings.LocalAuthDisabled,
+		LocalMfaEnabled:                 &settings.LocalMfaEnabled,
 	}
 
 	if settings.NetworkRange.IsValid() {
 		networkRangeStr := settings.NetworkRange.String()
 		apiSettings.NetworkRange = &networkRangeStr
 	}
+	if settings.NetworkRangeV6.IsValid() {
+		networkRangeV6Str := settings.NetworkRangeV6.String()
+		apiSettings.NetworkRangeV6 = &networkRangeV6Str
+	}
+	if settings.DashboardFeatures != nil {
+		apiSettings.DashboardFeatures = &api.AccountDashboardFeatures{
+			AgentNetwork: settings.DashboardFeatures.AgentNetwork,
+		}
+	}
 
 	apiOnboarding := api.AccountOnboarding{
 		OnboardingFlowPending: onboarding.OnboardingFlowPending,

management/server/http/handlers/accounts/accounts_handler_test.go

@@ -5,12 +5,14 @@ import (
 	"context"
 	"encoding/json"
 	"io"
+	"math"
 	"net/http"
 	"net/http/httptest"
+	"net/netip"
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 
@@ -32,6 +34,10 @@ func initAccountsTestData(t *testing.T, account *types.Account) *handler {
 		GetSettings(gomock.Any(), account.Id, "test_user").
 		Return(account.Settings, nil).
 		AnyTimes()
+	settingsMockManager.EXPECT().
+		GetEffectiveNetworkRanges(gomock.Any(), account.Id).
+		Return(netip.Prefix{}, netip.Prefix{}, nil).
+		AnyTimes()
 
 	return &handler{
 		accountManager: &mock_server.MockAccountManager{
@@ -124,8 +130,11 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 				DnsDomain:                       sr(""),
 				AutoUpdateAlways:                br(false),
 				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
 				EmbeddedIdpEnabled:              br(false),
 				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
 			},
 			expectedArray: true,
 			expectedID:    accountID,
@@ -150,8 +159,11 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 				DnsDomain:                       sr(""),
 				AutoUpdateAlways:                br(false),
 				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
 				EmbeddedIdpEnabled:              br(false),
 				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
 			},
 			expectedArray: false,
 			expectedID:    accountID,
@@ -176,8 +188,11 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 				DnsDomain:                       sr(""),
 				AutoUpdateAlways:                br(false),
 				AutoUpdateVersion:               sr("latest"),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
 				EmbeddedIdpEnabled:              br(false),
 				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
 			},
 			expectedArray: false,
 			expectedID:    accountID,
@@ -202,8 +217,11 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 				DnsDomain:                       sr(""),
 				AutoUpdateAlways:                br(false),
 				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
 				EmbeddedIdpEnabled:              br(false),
 				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
 			},
 			expectedArray: false,
 			expectedID:    accountID,
@@ -228,8 +246,11 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 				DnsDomain:                       sr(""),
 				AutoUpdateAlways:                br(false),
 				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
 				EmbeddedIdpEnabled:              br(false),
 				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
 			},
 			expectedArray: false,
 			expectedID:    accountID,
@@ -254,8 +275,113 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 				DnsDomain:                       sr(""),
 				AutoUpdateAlways:                br(false),
 				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
 				EmbeddedIdpEnabled:              br(false),
 				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
+			},
+			expectedArray: false,
+			expectedID:    accountID,
+		},
+		{
+			name:           "PutAccount OK enabling agent_network_only",
+			expectedBody:   true,
+			requestType:    http.MethodPut,
+			requestPath:    "/api/accounts/" + accountID,
+			requestBody:    bytes.NewBufferString("{\"settings\": {\"peer_login_expiration\": 15552000,\"peer_login_expiration_enabled\": true,\"agent_network_only\": true,\"dashboard_features\": {\"agent_network\": true}},\"onboarding\": {\"onboarding_flow_pending\": true,\"signup_form_pending\": true}}"),
+			expectedStatus: http.StatusOK,
+			expectedSettings: api.AccountSettings{
+				PeerLoginExpiration:             15552000,
+				PeerLoginExpirationEnabled:      true,
+				GroupsPropagationEnabled:        br(false),
+				JwtGroupsClaimName:              sr(""),
+				JwtGroupsEnabled:                br(false),
+				JwtAllowGroups:                  &[]string{},
+				RegularUsersViewBlocked:         false,
+				RoutingPeerDnsResolutionEnabled: br(false),
+				LazyConnectionEnabled:           br(false),
+				DnsDomain:                       sr(""),
+				AutoUpdateAlways:                br(false),
+				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(true),
+				DashboardFeatures: &api.AccountDashboardFeatures{
+					AgentNetwork: br(true),
+				},
+				EmbeddedIdpEnabled: br(false),
+				LocalAuthDisabled:  br(false),
+				LocalMfaEnabled:    br(false),
+			},
+			expectedArray: false,
+			expectedID:    accountID,
+		},
+		{
+			name:           "PutAccount fails enabling agent_network_only without dashboard_features",
+			expectedBody:   true,
+			requestType:    http.MethodPut,
+			requestPath:    "/api/accounts/" + accountID,
+			requestBody:    bytes.NewBufferString("{\"settings\": {\"peer_login_expiration\": 15552000,\"peer_login_expiration_enabled\": true,\"agent_network_only\": true},\"onboarding\": {\"onboarding_flow_pending\": true,\"signup_form_pending\": true}}"),
+			expectedStatus: http.StatusUnprocessableEntity,
+			expectedArray:  false,
+		},
+		{
+			name:           "PutAccount OK setting dashboard_features agent_network",
+			expectedBody:   true,
+			requestType:    http.MethodPut,
+			requestPath:    "/api/accounts/" + accountID,
+			requestBody:    bytes.NewBufferString("{\"settings\": {\"peer_login_expiration\": 15552000,\"peer_login_expiration_enabled\": true,\"dashboard_features\": {\"agent_network\": true}},\"onboarding\": {\"onboarding_flow_pending\": true,\"signup_form_pending\": true}}"),
+			expectedStatus: http.StatusOK,
+			expectedSettings: api.AccountSettings{
+				PeerLoginExpiration:             15552000,
+				PeerLoginExpirationEnabled:      true,
+				GroupsPropagationEnabled:        br(false),
+				JwtGroupsClaimName:              sr(""),
+				JwtGroupsEnabled:                br(false),
+				JwtAllowGroups:                  &[]string{},
+				RegularUsersViewBlocked:         false,
+				RoutingPeerDnsResolutionEnabled: br(false),
+				LazyConnectionEnabled:           br(false),
+				DnsDomain:                       sr(""),
+				AutoUpdateAlways:                br(false),
+				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
+				DashboardFeatures: &api.AccountDashboardFeatures{
+					AgentNetwork: br(true),
+				},
+				EmbeddedIdpEnabled: br(false),
+				LocalAuthDisabled:  br(false),
+				LocalMfaEnabled:    br(false),
+			},
+			expectedArray: false,
+			expectedID:    accountID,
+		},
+		{
+			name:           "PutAccount OK disabling agent_network_only again",
+			expectedBody:   true,
+			requestType:    http.MethodPut,
+			requestPath:    "/api/accounts/" + accountID,
+			requestBody:    bytes.NewBufferString("{\"settings\": {\"peer_login_expiration\": 15552000,\"peer_login_expiration_enabled\": true,\"agent_network_only\": false},\"onboarding\": {\"onboarding_flow_pending\": true,\"signup_form_pending\": true}}"),
+			expectedStatus: http.StatusOK,
+			expectedSettings: api.AccountSettings{
+				PeerLoginExpiration:             15552000,
+				PeerLoginExpirationEnabled:      true,
+				GroupsPropagationEnabled:        br(false),
+				JwtGroupsClaimName:              sr(""),
+				JwtGroupsEnabled:                br(false),
+				JwtAllowGroups:                  &[]string{},
+				RegularUsersViewBlocked:         false,
+				RoutingPeerDnsResolutionEnabled: br(false),
+				LazyConnectionEnabled:           br(false),
+				DnsDomain:                       sr(""),
+				AutoUpdateAlways:                br(false),
+				AutoUpdateVersion:               sr(""),
+				MetricsPushEnabled:              br(false),
+				AgentNetworkOnly:                br(false),
+				EmbeddedIdpEnabled:              br(false),
+				LocalAuthDisabled:               br(false),
+				LocalMfaEnabled:                 br(false),
 			},
 			expectedArray: false,
 			expectedID:    accountID,
@@ -337,3 +463,27 @@ func TestAccounts_AccountsHandler(t *testing.T) {
 		})
 	}
 }
+
+func TestCalculateMaxHosts(t *testing.T) {
+	tests := []struct {
+		name   string
+		prefix string
+		min    int64
+	}{
+		{"v4 /24", "100.64.0.0/24", 254},
+		{"v4 /16", "100.64.0.0/16", 65534},
+		{"v4 /28", "100.64.0.0/28", 14},
+		{"v6 /64", "fd00::/64", math.MaxInt64},
+		{"v6 /120", "fd00::/120", 256},
+		{"v6 /112", "fd00::/112", 65536},
+		{"v6 /48", "fd00::/48", math.MaxInt64},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			prefix := netip.MustParsePrefix(tt.prefix)
+			got := calculateMaxHosts(prefix)
+			assert.Equal(t, tt.min, got)
+		})
+	}
+}

management/server/http/handlers/dns/nameservers_handler.go

@@ -3,7 +3,10 @@ package dns
 import (
 	"encoding/json"
 	"fmt"
+	"net"
 	"net/http"
+	"strconv"
+	"strings"
 
 	"github.com/gorilla/mux"
 
@@ -162,7 +165,11 @@ func (h *nameserversHandler) getNameserverGroup(w http.ResponseWriter, r *http.R
 func toServerNSList(apiNSList []api.Nameserver) ([]nbdns.NameServer, error) {
 	var nsList []nbdns.NameServer
 	for _, apiNS := range apiNSList {
-		parsed, err := nbdns.ParseNameServerURL(fmt.Sprintf("%s://%s:%d", apiNS.NsType, apiNS.Ip, apiNS.Port))
+		host, err := unwrapBracketedHost(apiNS.Ip)
+		if err != nil {
+			return nil, err
+		}
+		parsed, err := nbdns.ParseNameServerURL(fmt.Sprintf("%s://%s", apiNS.NsType, net.JoinHostPort(host, strconv.Itoa(apiNS.Port))))
 		if err != nil {
 			return nil, err
 		}
@@ -172,6 +179,18 @@ func toServerNSList(apiNSList []api.Nameserver) ([]nbdns.NameServer, error) {
 	return nsList, nil
 }
 
+// unwrapBracketedHost returns ip with surrounding brackets stripped, rejecting
+// inputs with mismatched brackets.
+func unwrapBracketedHost(ip string) (string, error) {
+	if !strings.ContainsAny(ip, "[]") {
+		return ip, nil
+	}
+	if !strings.HasPrefix(ip, "[") || !strings.HasSuffix(ip, "]") {
+		return "", fmt.Errorf("malformed bracketed address: %s", ip)
+	}
+	return ip[1 : len(ip)-1], nil
+}
+
 func toNameserverGroupResponse(serverNSGroup *nbdns.NameServerGroup) *api.NameserverGroup {
 	var nsList []api.Nameserver
 	for _, ns := range serverNSGroup.NameServers {

management/server/http/handlers/dns/nameservers_handler_test.go

@@ -234,3 +234,37 @@ func TestNameserversHandlers(t *testing.T) {
 		})
 	}
 }
+
+func TestToServerNSList_IPv6(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    []api.Nameserver
+		expectIP netip.Addr
+	}{
+		{
+			name: "IPv4",
+			input: []api.Nameserver{
+				{Ip: "1.1.1.1", NsType: "udp", Port: 53},
+			},
+			expectIP: netip.MustParseAddr("1.1.1.1"),
+		},
+		{
+			name: "IPv6",
+			input: []api.Nameserver{
+				{Ip: "2001:4860:4860::8888", NsType: "udp", Port: 53},
+			},
+			expectIP: netip.MustParseAddr("2001:4860:4860::8888"),
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			result, err := toServerNSList(tc.input)
+			assert.NoError(t, err)
+			if assert.Len(t, result, 1) {
+				assert.Equal(t, tc.expectIP, result[0].IP)
+				assert.Equal(t, 53, result[0].Port)
+			}
+		})
+	}
+}

management/server/http/handlers/groups/groups_handler.go

@@ -43,7 +43,7 @@ func newHandler(accountManager account.Manager, permissionsManager permissions.M
 }
 
 func (h *handler) canReadPeers(r *http.Request, userAuth *auth.UserAuth) bool {
-	allowed, err := h.permissionsManager.ValidateUserPermissions(r.Context(), userAuth.AccountId, userAuth.UserId, modules.Peers, operations.Read)
+	allowed, _, err := h.permissionsManager.ValidateUserPermissions(r.Context(), userAuth.AccountId, userAuth.UserId, modules.Peers, operations.Read)
 	return err == nil && allowed
 }
 

management/server/http/handlers/groups/groups_handler_test.go

@@ -7,15 +7,15 @@ import (
 	"errors"
 	"fmt"
 	"io"
-	"net"
 	"net/http"
 	"net/http/httptest"
+	"net/netip"
 	"strings"
 	"testing"
 
-	"github.com/golang/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
+	"go.uber.org/mock/gomock"
 	"golang.org/x/exp/maps"
 
 	"github.com/netbirdio/netbird/management/internals/modules/permissions"
@@ -33,8 +33,8 @@ import (
 )
 
 var TestPeers = map[string]*nbpeer.Peer{
-	"A": {Key: "A", ID: "peer-A-ID", IP: net.ParseIP("100.100.100.100")},
-	"B": {Key: "B", ID: "peer-B-ID", IP: net.ParseIP("200.200.200.200")},
+	"A": {Key: "A", ID: "peer-A-ID", IP: netip.MustParseAddr("100.100.100.100")},
+	"B": {Key: "B", ID: "peer-B-ID", IP: netip.MustParseAddr("200.200.200.200")},
 }
 
 func initGroupTestData(t *testing.T, initGroups ...*types.Group) *handler {
@@ -44,7 +44,7 @@ func initGroupTestData(t *testing.T, initGroups ...*types.Group) *handler {
 	permissionsManagerMock := permissions.NewMockManager(ctrl)
 	permissionsManagerMock.EXPECT().
 		ValidateUserPermissions(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Eq(modules.Peers), gomock.Eq(operations.Read)).
-		Return(true, nil).
+		Return(true, context.Background(), nil).
 		AnyTimes()
 
 	return &handler{

management/server/http/handlers/instance/instance_handler.go

@@ -10,6 +10,7 @@ import (
 	"github.com/netbirdio/netbird/management/internals/modules/permissions"
 	"github.com/netbirdio/netbird/management/internals/modules/permissions/modules"
 	"github.com/netbirdio/netbird/management/internals/modules/permissions/operations"
+	"github.com/netbirdio/netbird/management/server/account"
 	nbinstance "github.com/netbirdio/netbird/management/server/instance"
 	"github.com/netbirdio/netbird/shared/auth"
 	"github.com/netbirdio/netbird/shared/management/http/api"
@@ -19,13 +20,15 @@ import (
 // handler handles the instance setup HTTP endpoints
 type handler struct {
 	instanceManager nbinstance.Manager
+	setupManager    *nbinstance.SetupService
 }
 
 // AddEndpoints registers the instance setup endpoints.
 // These endpoints bypass authentication for initial setup.
-func AddEndpoints(instanceManager nbinstance.Manager, router *mux.Router) {
+func AddEndpoints(instanceManager nbinstance.Manager, accountManager account.Manager, router *mux.Router) {
 	h := &handler{
 		instanceManager: instanceManager,
+		setupManager:    nbinstance.NewSetupService(instanceManager, accountManager),
 	}
 
 	router.HandleFunc("/instance", h.getInstanceStatus).Methods("GET", "OPTIONS")
@@ -59,24 +62,36 @@ func (h *handler) getInstanceStatus(w http.ResponseWriter, r *http.Request) {
 // setup creates the initial admin user for the instance.
 // This endpoint is unauthenticated but only works when setup is required.
 func (h *handler) setup(w http.ResponseWriter, r *http.Request) {
+	ctx := r.Context()
+
 	var req api.SetupRequest
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
 		util.WriteErrorResponse("invalid request body", http.StatusBadRequest, w)
 		return
 	}
 
-	userData, err := h.instanceManager.CreateOwnerUser(r.Context(), req.Email, req.Password, req.Name)
+	result, err := h.setupManager.SetupOwner(ctx, req.Email, req.Password, req.Name, nbinstance.SetupOptions{
+		CreatePAT:       req.CreatePat != nil && *req.CreatePat,
+		PATExpireInDays: req.PatExpireIn,
+	})
 	if err != nil {
-		util.WriteError(r.Context(), err, w)
+		util.WriteError(ctx, err, w)
 		return
 	}
 
-	log.WithContext(r.Context()).Infof("instance setup completed: created user %s", req.Email)
+	log.WithContext(ctx).Infof("instance setup completed: created user %s", req.Email)
 
-	util.WriteJSONObject(r.Context(), w, api.SetupResponse{
-		UserId: userData.ID,
-		Email:  userData.Email,
-	})
+	resp := api.SetupResponse{
+		UserId: result.User.ID,
+		Email:  result.User.Email,
+	}
+
+	if result.PATPlainToken != "" {
+		resp.PersonalAccessToken = &result.PATPlainToken
+	}
+
+	w.Header().Set("Cache-Control", "no-store")
+	util.WriteJSONObject(ctx, w, resp)
 }
 
 // getVersionInfo returns version information for NetBird components.

management/server/http/handlers/instance/instance_handler_test.go

@@ -10,16 +10,20 @@ import (
 	"net/mail"
 	"testing"
 
-	"github.com/golang/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+	"go.uber.org/mock/gomock"
 
 	"github.com/netbirdio/netbird/management/internals/modules/permissions"
 	"github.com/netbirdio/netbird/management/internals/modules/permissions/modules"
 	"github.com/netbirdio/netbird/management/internals/modules/permissions/operations"
+	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/idp"
 	nbinstance "github.com/netbirdio/netbird/management/server/instance"
+	"github.com/netbirdio/netbird/management/server/mock_server"
+	nbstore "github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/shared/auth"
 	"github.com/netbirdio/netbird/shared/management/http/api"
 	"github.com/netbirdio/netbird/shared/management/status"
@@ -30,6 +34,7 @@ type mockInstanceManager struct {
 	isSetupRequired   bool
 	isSetupRequiredFn func(ctx context.Context) (bool, error)
 	createOwnerUserFn func(ctx context.Context, email, password, name string) (*idp.UserData, error)
+	rollbackSetupFn   func(ctx context.Context, userID string) error
 	getVersionInfoFn  func(ctx context.Context) (*nbinstance.VersionInfo, error)
 }
 
@@ -72,6 +77,13 @@ func (m *mockInstanceManager) CreateOwnerUser(ctx context.Context, email, passwo
 	}, nil
 }
 
+func (m *mockInstanceManager) RollbackSetup(ctx context.Context, userID string) error {
+	if m.rollbackSetupFn != nil {
+		return m.rollbackSetupFn(ctx, userID)
+	}
+	return nil
+}
+
 func (m *mockInstanceManager) GetVersionInfo(ctx context.Context) (*nbinstance.VersionInfo, error) {
 	if m.getVersionInfoFn != nil {
 		return m.getVersionInfoFn(ctx)
@@ -87,8 +99,12 @@ func (m *mockInstanceManager) GetVersionInfo(ctx context.Context) (*nbinstance.V
 var _ nbinstance.Manager = (*mockInstanceManager)(nil)
 
 func setupTestRouter(manager nbinstance.Manager) *mux.Router {
+	return setupTestRouterWithPAT(manager, nil)
+}
+
+func setupTestRouterWithPAT(manager nbinstance.Manager, accountManager account.Manager) *mux.Router {
 	router := mux.NewRouter()
-	AddEndpoints(manager, router)
+	AddEndpoints(manager, accountManager, router)
 	return router
 }
 
@@ -166,6 +182,7 @@ func TestSetup_Success(t *testing.T) {
 	router.ServeHTTP(rec, req)
 
 	assert.Equal(t, http.StatusOK, rec.Code)
+	assert.Equal(t, "no-store", rec.Header().Get("Cache-Control"))
 
 	var response api.SetupResponse
 	err := json.NewDecoder(rec.Body).Decode(&response)
@@ -298,6 +315,239 @@ func TestSetup_ManagerError(t *testing.T) {
 	assert.Equal(t, http.StatusInternalServerError, rec.Code)
 }
 
+func TestSetup_PAT_FeatureDisabled_IgnoresCreatePAT(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "false")
+
+	manager := &mockInstanceManager{isSetupRequired: true}
+	// NB_SETUP_PAT_ENABLED=false: request fields must be silently ignored
+	router := setupTestRouterWithPAT(manager, &mock_server.MockAccountManager{})
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin", "create_pat": true}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusOK, rec.Code)
+	var response api.SetupResponse
+	require.NoError(t, json.NewDecoder(rec.Body).Decode(&response))
+	assert.Nil(t, response.PersonalAccessToken)
+}
+
+func TestSetup_PAT_FlagOmitted_NoPAT(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "true")
+
+	manager := &mockInstanceManager{isSetupRequired: true}
+	router := setupTestRouterWithPAT(manager, &mock_server.MockAccountManager{})
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin"}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusOK, rec.Code)
+	var response api.SetupResponse
+	require.NoError(t, json.NewDecoder(rec.Body).Decode(&response))
+	assert.Nil(t, response.PersonalAccessToken)
+}
+
+func TestSetup_PAT_MissingExpireIn_DefaultsToOneDay(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "true")
+
+	createCalled := false
+	manager := &mockInstanceManager{
+		isSetupRequired: true,
+		createOwnerUserFn: func(ctx context.Context, email, password, name string) (*idp.UserData, error) {
+			createCalled = true
+			return &idp.UserData{ID: "u1", Email: email, Name: name}, nil
+		},
+	}
+	accountMgr := &mock_server.MockAccountManager{
+		GetAccountIDByUserIdFunc: func(_ context.Context, userAuth auth.UserAuth) (string, error) {
+			assert.Equal(t, "u1", userAuth.UserId)
+			return "acc-1", nil
+		},
+		CreatePATFunc: func(_ context.Context, accountID, initiator, target, name string, expiresIn int) (*types.PersonalAccessTokenGenerated, error) {
+			assert.Equal(t, "acc-1", accountID)
+			assert.Equal(t, "u1", initiator)
+			assert.Equal(t, "u1", target)
+			assert.Equal(t, "setup-token", name)
+			assert.Equal(t, 1, expiresIn)
+			return &types.PersonalAccessTokenGenerated{PlainToken: "nbp_plain"}, nil
+		},
+	}
+	router := setupTestRouterWithPAT(manager, accountMgr)
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin", "create_pat": true}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusOK, rec.Code)
+	assert.Equal(t, "no-store", rec.Header().Get("Cache-Control"))
+	assert.True(t, createCalled)
+	var response api.SetupResponse
+	require.NoError(t, json.NewDecoder(rec.Body).Decode(&response))
+	require.NotNil(t, response.PersonalAccessToken)
+	assert.Equal(t, "nbp_plain", *response.PersonalAccessToken)
+}
+
+func TestSetup_PAT_ExpireOutOfRange(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "true")
+
+	manager := &mockInstanceManager{isSetupRequired: true}
+	router := setupTestRouterWithPAT(manager, &mock_server.MockAccountManager{})
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin", "create_pat": true, "pat_expire_in": 0}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusUnprocessableEntity, rec.Code)
+}
+
+func TestSetup_PAT_Success(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "true")
+
+	manager := &mockInstanceManager{
+		isSetupRequired: true,
+		createOwnerUserFn: func(ctx context.Context, email, password, name string) (*idp.UserData, error) {
+			return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+		},
+	}
+
+	gotAccountArgs := struct {
+		userID string
+		email  string
+	}{}
+	accountMgr := &mock_server.MockAccountManager{
+		GetAccountIDByUserIdFunc: func(_ context.Context, userAuth auth.UserAuth) (string, error) {
+			gotAccountArgs.userID = userAuth.UserId
+			gotAccountArgs.email = userAuth.Email
+			return "acc-1", nil
+		},
+		CreatePATFunc: func(_ context.Context, accountID, initiator, target, name string, expiresIn int) (*types.PersonalAccessTokenGenerated, error) {
+			assert.Equal(t, "acc-1", accountID)
+			assert.Equal(t, "owner-id", initiator)
+			assert.Equal(t, "owner-id", target)
+			assert.Equal(t, "setup-token", name)
+			assert.Equal(t, 30, expiresIn)
+			return &types.PersonalAccessTokenGenerated{PlainToken: "nbp_plain"}, nil
+		},
+	}
+
+	router := setupTestRouterWithPAT(manager, accountMgr)
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin", "create_pat": true, "pat_expire_in": 30}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusOK, rec.Code)
+	assert.Equal(t, "no-store", rec.Header().Get("Cache-Control"))
+	var response api.SetupResponse
+	require.NoError(t, json.NewDecoder(rec.Body).Decode(&response))
+	assert.Equal(t, "owner-id", response.UserId)
+	require.NotNil(t, response.PersonalAccessToken)
+	assert.Equal(t, "nbp_plain", *response.PersonalAccessToken)
+	assert.Equal(t, "owner-id", gotAccountArgs.userID)
+	assert.Equal(t, "admin@example.com", gotAccountArgs.email)
+}
+
+func TestSetup_PAT_AccountCreationFails_Rollback(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "true")
+
+	ctrl := gomock.NewController(t)
+	accountStore := nbstore.NewMockStore(ctrl)
+	accountStore.EXPECT().GetAccountIDByUserID(gomock.Any(), nbstore.LockingStrengthNone, "owner-id").Return("", status.NewAccountNotFoundError("owner-id"))
+
+	rolledBackFor := ""
+	manager := &mockInstanceManager{
+		isSetupRequired: true,
+		createOwnerUserFn: func(ctx context.Context, email, password, name string) (*idp.UserData, error) {
+			return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+		},
+		rollbackSetupFn: func(_ context.Context, userID string) error {
+			rolledBackFor = userID
+			return nil
+		},
+	}
+	accountMgr := &mock_server.MockAccountManager{
+		GetAccountIDByUserIdFunc: func(_ context.Context, _ auth.UserAuth) (string, error) {
+			return "", errors.New("db down")
+		},
+		GetStoreFunc: func() nbstore.Store {
+			return accountStore
+		},
+	}
+
+	router := setupTestRouterWithPAT(manager, accountMgr)
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin", "create_pat": true, "pat_expire_in": 30}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusInternalServerError, rec.Code)
+	assert.Equal(t, "owner-id", rolledBackFor, "RollbackSetup must be called with the created user id")
+}
+
+func TestSetup_PAT_CreatePATFails_Rollback(t *testing.T) {
+	t.Setenv(nbinstance.SetupPATEnabledEnvKey, "true")
+
+	ctrl := gomock.NewController(t)
+	accountStore := nbstore.NewMockStore(ctrl)
+	account := &types.Account{Id: "acc-1"}
+	accountStore.EXPECT().GetAccount(gomock.Any(), "acc-1").Return(account, nil)
+	accountStore.EXPECT().DeleteAccount(gomock.Any(), account).Return(nil)
+
+	rolledBackFor := ""
+	manager := &mockInstanceManager{
+		isSetupRequired: true,
+		createOwnerUserFn: func(ctx context.Context, email, password, name string) (*idp.UserData, error) {
+			return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+		},
+		rollbackSetupFn: func(_ context.Context, userID string) error {
+			rolledBackFor = userID
+			return nil
+		},
+	}
+	accountMgr := &mock_server.MockAccountManager{
+		GetAccountIDByUserIdFunc: func(_ context.Context, _ auth.UserAuth) (string, error) {
+			return "acc-1", nil
+		},
+		CreatePATFunc: func(_ context.Context, _, _, _, _ string, _ int) (*types.PersonalAccessTokenGenerated, error) {
+			return nil, status.Errorf(status.Internal, "token store unavailable")
+		},
+		GetStoreFunc: func() nbstore.Store {
+			return accountStore
+		},
+	}
+
+	router := setupTestRouterWithPAT(manager, accountMgr)
+
+	body := `{"email": "admin@example.com", "password": "securepassword123", "name": "Admin", "create_pat": true, "pat_expire_in": 30}`
+	req := httptest.NewRequest(http.MethodPost, "/setup", bytes.NewBufferString(body))
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	router.ServeHTTP(rec, req)
+
+	assert.Equal(t, http.StatusInternalServerError, rec.Code)
+	assert.Equal(t, "owner-id", rolledBackFor, "RollbackSetup must be called when CreatePAT fails")
+}
+
 func TestGetVersionInfo_Success(t *testing.T) {
 	manager := &mockInstanceManager{}
 	ctrl := gomock.NewController(t)

management/server/http/handlers/peers/peers_handler.go

@@ -214,6 +214,18 @@ func (h *Handler) UpdatePeer(w http.ResponseWriter, r *http.Request, userAuth *a
 		}
 	}
 
+	if req.Ipv6 != nil {
+		v6Addr, err := parseIPv6(req.Ipv6)
+		if err != nil {
+			util.WriteError(r.Context(), status.Errorf(status.InvalidArgument, "%v", err), w)
+			return
+		}
+		if err = h.accountManager.UpdatePeerIPv6(r.Context(), userAuth.AccountId, userAuth.UserId, peerID, v6Addr); err != nil {
+			util.WriteError(r.Context(), err, w)
+			return
+		}
+	}
+
 	peer, err := h.accountManager.UpdatePeer(r.Context(), userAuth.AccountId, userAuth.UserId, update)
 	if err != nil {
 		util.WriteError(r.Context(), err, w)
@@ -318,6 +330,21 @@ func (h *Handler) setApprovalRequiredFlag(respBody []*api.PeerBatch, validPeersM
 	}
 }
 
+func parseIPv6(s *string) (netip.Addr, error) {
+	if s == nil {
+		return netip.Addr{}, fmt.Errorf("IPv6 address is nil")
+	}
+	addr, err := netip.ParseAddr(*s)
+	if err != nil {
+		return netip.Addr{}, fmt.Errorf("invalid IPv6 address %s: %w", *s, err)
+	}
+	addr = addr.Unmap()
+	if !addr.Is6() {
+		return netip.Addr{}, fmt.Errorf("address %s is not IPv6", *s)
+	}
+	return addr, nil
+}
+
 // GetAccessiblePeers returns a list of all peers that the specified peer can connect to within the network.
 func (h *Handler) GetAccessiblePeers(w http.ResponseWriter, r *http.Request, userAuth *auth.UserAuth) {
 	vars := mux.Vars(r)
@@ -369,9 +396,9 @@ func (h *Handler) GetAccessiblePeers(w http.ResponseWriter, r *http.Request, use
 
 	dnsDomain := h.networkMapController.GetDNSDomain(account.Settings)
 
-	netMap := account.GetPeerNetworkMap(r.Context(), peerID, dns.CustomZone{}, nil, validPeers, account.GetResourcePoliciesMap(), account.GetResourceRoutersMap(), nil, account.GetActiveGroupUsers())
+	netMap := account.GetPeerNetworkMapFromComponents(r.Context(), peerID, dns.CustomZone{}, nil, validPeers, account.GetResourcePoliciesMap(), account.GetResourceRoutersMap(), nil, account.GetActiveGroupUsers())
 
-	util.WriteJSONObject(r.Context(), w, toAccessiblePeers(netMap, dnsDomain))
+	util.WriteJSONObject(r.Context(), w, toAccessiblePeers(account.Peers, netMap, dnsDomain))
 }
 
 func (h *Handler) CreateTemporaryAccess(w http.ResponseWriter, r *http.Request, userAuth *auth.UserAuth) {
@@ -398,7 +425,7 @@ func (h *Handler) CreateTemporaryAccess(w http.ResponseWriter, r *http.Request,
 		return
 	}
 
-	peer, _, _, err := h.accountManager.AddPeer(r.Context(), userAuth.AccountId, "", userAuth.UserId, newPeer, true)
+	peer, _, _, _, err := h.accountManager.AddPeer(r.Context(), userAuth.AccountId, "", userAuth.UserId, newPeer, true)
 	if err != nil {
 		util.WriteError(r.Context(), err, w)
 		return
@@ -453,14 +480,21 @@ func (h *Handler) CreateTemporaryAccess(w http.ResponseWriter, r *http.Request,
 	util.WriteJSONObject(r.Context(), w, resp)
 }
 
-func toAccessiblePeers(netMap *types.NetworkMap, dnsDomain string) []api.AccessiblePeer {
+// toAccessiblePeers resolves the twin peers in netMap back to the full account
+// peers (by ID) so the API response keeps Status/Name/OS/GeoNameID, which the
+// slim netmap twins intentionally don't carry.
+func toAccessiblePeers(accountPeers map[string]*nbpeer.Peer, netMap *types.NetworkMap, dnsDomain string) []api.AccessiblePeer {
 	accessiblePeers := make([]api.AccessiblePeer, 0, len(netMap.Peers)+len(netMap.OfflinePeers))
+	appendByID := func(id string) {
+		if p, ok := accountPeers[id]; ok && p != nil {
+			accessiblePeers = append(accessiblePeers, peerToAccessiblePeer(p, dnsDomain))
+		}
+	}
 	for _, p := range netMap.Peers {
-		accessiblePeers = append(accessiblePeers, peerToAccessiblePeer(p, dnsDomain))
+		appendByID(p.ID)
 	}
-
 	for _, p := range netMap.OfflinePeers {
-		accessiblePeers = append(accessiblePeers, peerToAccessiblePeer(p, dnsDomain))
+		appendByID(p.ID)
 	}
 
 	return accessiblePeers
@@ -475,6 +509,7 @@ func peerToAccessiblePeer(peer *nbpeer.Peer, dnsDomain string) api.AccessiblePee
 		GeonameId:   int(peer.Location.GeoNameID),
 		Id:          peer.ID,
 		Ip:          peer.IP.String(),
+		Ipv6:        peerIPv6String(peer),
 		LastSeen:    peer.Status.LastSeen,
 		Name:        peer.Name,
 		Os:          peer.Meta.OS,
@@ -493,6 +528,7 @@ func toSinglePeerResponse(peer *nbpeer.Peer, groupsInfo []api.GroupMinimum, dnsD
 		Id:                          peer.ID,
 		Name:                        peer.Name,
 		Ip:                          peer.IP.String(),
+		Ipv6:                        peerIPv6String(peer),
 		ConnectionIp:                peer.Location.ConnectionIP.String(),
 		Connected:                   peer.Status.Connected,
 		LastSeen:                    peer.Status.LastSeen,
@@ -527,6 +563,7 @@ func toSinglePeerResponse(peer *nbpeer.Peer, groupsInfo []api.GroupMinimum, dnsD
 			RosenpassEnabled:      &peer.Meta.Flags.RosenpassEnabled,
 			RosenpassPermissive:   &peer.Meta.Flags.RosenpassPermissive,
 			ServerSshAllowed:      &peer.Meta.Flags.ServerSSHAllowed,
+			RemoteJobsAllowed:     &peer.Meta.Flags.RemoteJobsAllowed,
 		},
 	}
 
@@ -547,6 +584,7 @@ func toPeerListItemResponse(peer *nbpeer.Peer, groupsInfo []api.GroupMinimum, dn
 		Id:                          peer.ID,
 		Name:                        peer.Name,
 		Ip:                          peer.IP.String(),
+		Ipv6:                        peerIPv6String(peer),
 		ConnectionIp:                peer.Location.ConnectionIP.String(),
 		Connected:                   peer.Status.Connected,
 		LastSeen:                    peer.Status.LastSeen,
@@ -581,6 +619,7 @@ func toPeerListItemResponse(peer *nbpeer.Peer, groupsInfo []api.GroupMinimum, dn
 			RosenpassEnabled:      &peer.Meta.Flags.RosenpassEnabled,
 			RosenpassPermissive:   &peer.Meta.Flags.RosenpassPermissive,
 			ServerSshAllowed:      &peer.Meta.Flags.ServerSSHAllowed,
+			RemoteJobsAllowed:     &peer.Meta.Flags.RemoteJobsAllowed,
 		},
 	}
 }
@@ -623,3 +662,11 @@ func fqdnList(extraLabels []string, dnsDomain string) []string {
 	}
 	return fqdnList
 }
+
+func peerIPv6String(peer *nbpeer.Peer) *string {
+	if !peer.IPv6.IsValid() {
+		return nil
+	}
+	s := peer.IPv6.String()
+	return &s
+}

management/server/http/handlers/peers/peers_handler_test.go

@@ -13,9 +13,8 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
 	"github.com/gorilla/mux"
-	ugomock "go.uber.org/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"golang.org/x/exp/maps"
 
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map"
@@ -106,7 +105,7 @@ func initTestMetaData(t *testing.T, peers ...*nbpeer.Peer) *Handler {
 		},
 	}
 
-	ctrl := ugomock.NewController(t)
+	ctrl := gomock.NewController(t)
 
 	networkMapController := network_map.NewMockController(ctrl)
 	networkMapController.EXPECT().
@@ -116,15 +115,15 @@ func initTestMetaData(t *testing.T, peers ...*nbpeer.Peer) *Handler {
 
 	ctrl2 := gomock.NewController(t)
 	permissionsManager := permissions.NewMockManager(ctrl2)
-	permissionsManager.EXPECT().ValidateAccountAccess(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
+	permissionsManager.EXPECT().ValidateAccountAccess(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).Return(context.Background(), nil).AnyTimes()
 	permissionsManager.EXPECT().
 		ValidateUserPermissions(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Eq(modules.Peers), gomock.Eq(operations.Read)).
-		DoAndReturn(func(ctx context.Context, accountID, userID string, module modules.Module, operation operations.Operation) (bool, error) {
+		DoAndReturn(func(ctx context.Context, accountID, userID string, module modules.Module, operation operations.Operation) (bool, context.Context, error) {
 			user, ok := account.Users[userID]
 			if !ok {
-				return false, fmt.Errorf("user not found")
+				return false, ctx, fmt.Errorf("user not found")
 			}
-			return user.HasAdminPower() || user.IsServiceUser, nil
+			return user.HasAdminPower() || user.IsServiceUser, ctx, nil
 		}).
 		AnyTimes()
 
@@ -146,7 +145,7 @@ func initTestMetaData(t *testing.T, peers ...*nbpeer.Peer) *Handler {
 			UpdatePeerIPFunc: func(_ context.Context, accountID, userID, peerID string, newIP netip.Addr) error {
 				for _, peer := range peers {
 					if peer.ID == peerID {
-						peer.IP = net.IP(newIP.AsSlice())
+						peer.IP = newIP
 						return nil
 					}
 				}
@@ -228,7 +227,8 @@ func TestGetPeers(t *testing.T) {
 	peer := &nbpeer.Peer{
 		ID:                     testPeerID,
 		Key:                    "key",
-		IP:                     net.ParseIP("100.64.0.1"),
+		IP:                     netip.MustParseAddr("100.64.0.1"),
+		IPv6:                   netip.MustParseAddr("fd00::1"),
 		Status:                 &nbpeer.PeerStatus{Connected: true},
 		Name:                   "PeerName",
 		LoginExpirationEnabled: false,
@@ -368,7 +368,8 @@ func TestGetAccessiblePeers(t *testing.T) {
 	peer1 := &nbpeer.Peer{
 		ID:                     "peer1",
 		Key:                    "key1",
-		IP:                     net.ParseIP("100.64.0.1"),
+		IP:                     netip.MustParseAddr("100.64.0.1"),
+		IPv6:                   netip.MustParseAddr("fd00:1234::1"),
 		Status:                 &nbpeer.PeerStatus{Connected: true},
 		Name:                   "peer1",
 		LoginExpirationEnabled: false,
@@ -378,7 +379,8 @@ func TestGetAccessiblePeers(t *testing.T) {
 	peer2 := &nbpeer.Peer{
 		ID:                     "peer2",
 		Key:                    "key2",
-		IP:                     net.ParseIP("100.64.0.2"),
+		IP:                     netip.MustParseAddr("100.64.0.2"),
+		IPv6:                   netip.MustParseAddr("fd00:1234::2"),
 		Status:                 &nbpeer.PeerStatus{Connected: true},
 		Name:                   "peer2",
 		LoginExpirationEnabled: false,
@@ -388,7 +390,8 @@ func TestGetAccessiblePeers(t *testing.T) {
 	peer3 := &nbpeer.Peer{
 		ID:                     "peer3",
 		Key:                    "key3",
-		IP:                     net.ParseIP("100.64.0.3"),
+		IP:                     netip.MustParseAddr("100.64.0.3"),
+		IPv6:                   netip.MustParseAddr("fd00:1234::3"),
 		Status:                 &nbpeer.PeerStatus{Connected: true},
 		Name:                   "peer3",
 		LoginExpirationEnabled: false,
@@ -532,7 +535,8 @@ func TestPeersHandlerUpdatePeerIP(t *testing.T) {
 	testPeer := &nbpeer.Peer{
 		ID:                     testPeerID,
 		Key:                    "key",
-		IP:                     net.ParseIP("100.64.0.1"),
+		IP:                     netip.MustParseAddr("100.64.0.1"),
+		IPv6:                   netip.MustParseAddr("fd00::1"),
 		Status:                 &nbpeer.PeerStatus{Connected: false, LastSeen: time.Now()},
 		Name:                   "test-host@netbird.io",
 		LoginExpirationEnabled: false,

management/server/http/handlers/policies/geolocation_handler_test.go

@@ -10,7 +10,7 @@ import (
 	"path/filepath"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 
@@ -51,7 +51,7 @@ func initGeolocationTestData(t *testing.T) *geolocationsHandler {
 	permissionsManagerMock.
 		EXPECT().
 		ValidateUserPermissions(gomock.Any(), gomock.Any(), gomock.Any(), modules.Policies, operations.Read).
-		Return(true, nil).
+		Return(true, context.Background(), nil).
 		AnyTimes()
 
 	return &geolocationsHandler{

management/server/http/handlers/proxy/auth.go

@@ -2,6 +2,7 @@ package proxy
 
 import (
 	"context"
+	"errors"
 	"net"
 	"net/http"
 	"net/netip"
@@ -99,16 +100,17 @@ func (h *AuthCallbackHandler) handleCallback(w http.ResponseWriter, r *http.Requ
 		return
 	}
 
-	// Group validation is performed by the proxy via ValidateSession gRPC call.
-	// This allows the proxy to show 403 pages directly without redirect dance.
-
+	// GenerateSessionToken applies the service's group and account-status gates,
+	// so a user without access never receives a token. The proxy re-checks the
+	// installed cookie against the service's allowed groups, and renders the
+	// denial page from the error carried back in the redirect.
 	sessionToken, err := h.proxyService.GenerateSessionToken(r.Context(), redirectURL.Hostname(), userID, auth.MethodOIDC)
 	if err != nil {
 		log.WithError(err).Error("Failed to create session token")
 		redirectURL.Scheme = "https"
 		query := redirectURL.Query()
 		query.Set("error", "access_denied")
-		query.Set("error_description", "Service configuration error")
+		query.Set("error_description", sessionTokenErrorDescription(err))
 		redirectURL.RawQuery = query.Encode()
 		http.Redirect(w, r, redirectURL.String(), http.StatusFound)
 		return
@@ -124,6 +126,23 @@ func (h *AuthCallbackHandler) handleCallback(w http.ResponseWriter, r *http.Requ
 	http.Redirect(w, r, redirectURL.String(), http.StatusFound)
 }
 
+// sessionTokenErrorDescription maps a session token failure to the text the
+// proxy renders on its access denied page. Account status denials get a message
+// the user can act on, while everything else stays generic so a lookup or
+// signing failure does not describe management internals to the browser.
+func sessionTokenErrorDescription(err error) string {
+	if errors.Is(err, nbgrpc.ErrUserPendingApproval) {
+		return "Your account is pending approval by an administrator"
+	}
+	if errors.Is(err, nbgrpc.ErrUserBlocked) {
+		return "Your account is blocked"
+	}
+	if errors.Is(err, nbgrpc.ErrUserNotInGroup) {
+		return "You are not authorized to access this service"
+	}
+	return "Service configuration error"
+}
+
 func extractUserIDFromToken(ctx context.Context, provider *oidc.Provider, config nbgrpc.ProxyOIDCConfig, token *oauth2.Token) string {
 	rawIDToken, ok := token.Extra("id_token").(string)
 	if !ok {

management/server/http/handlers/proxy/auth_callback_integration_test.go

@@ -19,6 +19,7 @@ import (
 	"github.com/stretchr/testify/require"
 
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/accesslogs"
+	activitymanager "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/activity/manager"
 	nbproxy "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
@@ -216,9 +217,12 @@ func setupAuthCallbackTest(t *testing.T) *testSetup {
 		nil,
 		usersManager,
 		nil,
+		nil,
+		nil,
 	)
 
 	proxyService.SetServiceManager(&testServiceManager{store: testStore})
+	proxyService.SetActivityManager(activitymanager.NewManager(testStore))
 
 	handler := NewAuthCallbackHandler(proxyService, nil)
 
@@ -358,6 +362,51 @@ func createTestAccountsAndUsers(t *testing.T, ctx context.Context, testStore sto
 		Issued:     "api",
 	}
 	require.NoError(t, testStore.SaveUser(ctx, allowedUser))
+
+	// A second tenant, whose users must never be issued a token signed with
+	// the first tenant's service session key.
+	otherAccount := &types.Account{
+		Id:                     "otherAccountId",
+		Domain:                 "other.com",
+		DomainCategory:         "private",
+		IsDomainPrimaryAccount: true,
+		CreatedAt:              time.Now(),
+	}
+	require.NoError(t, testStore.SaveAccount(ctx, otherAccount))
+
+	otherAccountUser := &types.User{
+		Id:        "otherAccountUserId",
+		AccountID: "otherAccountId",
+		Role:      types.UserRoleUser,
+		CreatedAt: time.Now(),
+		Issued:    "api",
+	}
+	require.NoError(t, testStore.SaveUser(ctx, otherAccountUser))
+
+	// A user awaiting approval is stored as blocked and pending approval, and
+	// carries the same group membership as the approved one.
+	pendingUser := &types.User{
+		Id:              "pendingUserId",
+		AccountID:       "testAccountId",
+		Role:            types.UserRoleUser,
+		AutoGroups:      []string{"allowedGroupId"},
+		Blocked:         true,
+		PendingApproval: true,
+		CreatedAt:       time.Now(),
+		Issued:          "api",
+	}
+	require.NoError(t, testStore.SaveUser(ctx, pendingUser))
+
+	blockedUser := &types.User{
+		Id:         "blockedUserId",
+		AccountID:  "testAccountId",
+		Role:       types.UserRoleUser,
+		AutoGroups: []string{"allowedGroupId"},
+		Blocked:    true,
+		CreatedAt:  time.Now(),
+		Issued:     "api",
+	}
+	require.NoError(t, testStore.SaveUser(ctx, blockedUser))
 }
 
 // testServiceManager is a minimal implementation for testing.
@@ -389,6 +438,10 @@ func (m *testServiceManager) DeleteService(_ context.Context, _, _, _ string) er
 	return nil
 }
 
+func (m *testServiceManager) DeleteAccountCluster(_ context.Context, _, _, _ string) error {
+	return nil
+}
+
 func (m *testServiceManager) SetCertificateIssuedAt(_ context.Context, _, _ string) error {
 	return nil
 }
@@ -435,7 +488,11 @@ func (m *testServiceManager) StopServiceFromPeer(_ context.Context, _, _, _ stri
 
 func (m *testServiceManager) StartExposeReaper(_ context.Context) {}
 
-func (m *testServiceManager) GetActiveClusters(_ context.Context, _, _ string) ([]nbproxy.Cluster, error) {
+func (m *testServiceManager) GetServiceByDomain(ctx context.Context, domain string) (*service.Service, error) {
+	return m.store.GetServiceByDomain(ctx, domain)
+}
+
+func (m *testServiceManager) GetClusters(_ context.Context, _, _ string) ([]nbproxy.Cluster, error) {
 	return nil, nil
 }
 
@@ -480,6 +537,113 @@ func TestAuthCallback_UserAllowedToLogin(t *testing.T) {
 	require.Empty(t, parsedLocation.Query().Get("error"), "Should not have error parameter")
 }
 
+// TestAuthCallback_UserDeniedByAccountStatus asserts that a user whose account
+// is pending approval or blocked never receives a session token from the OIDC
+// callback, and that the redirect carries a description the proxy can render.
+// TestAuthCallback_RecordsUserLogin drives the real OIDC callback and asserts
+// the login lands on the user row. That timestamp is what activity accounting
+// reads, and it is the only signal that can ever count someone who reaches
+// proxy-protected services from a browser and never opens the dashboard.
+func TestAuthCallback_RecordsUserLogin(t *testing.T) {
+	setup := setupAuthCallbackTest(t)
+	defer setup.cleanup()
+
+	ctx := context.Background()
+
+	before, err := setup.store.GetUserByUserID(ctx, store.LockingStrengthNone, "allowedUserId")
+	require.NoError(t, err)
+	require.Nil(t, before.LastLogin, "fixture user starts with no login on record")
+
+	setup.oidcServer.tokenSubject = "allowedUserId"
+	state := createTestState(t, setup.proxyService, "https://test-proxy.example.com/dashboard")
+
+	req := httptest.NewRequest(http.MethodGet, "/reverse-proxy/callback?code=test-auth-code&state="+url.QueryEscape(state), nil)
+	rec := httptest.NewRecorder()
+	setup.router.ServeHTTP(rec, req)
+	require.Equal(t, http.StatusFound, rec.Code)
+
+	after, err := setup.store.GetUserByUserID(ctx, store.LockingStrengthNone, "allowedUserId")
+	require.NoError(t, err)
+	require.NotNil(t, after.LastLogin, "a completed proxy SSO login must be recorded on the user")
+	require.WithinDuration(t, time.Now().UTC(), after.LastLogin.UTC(), time.Minute, "login should be stamped at sign-in time")
+}
+
+// TestAuthCallback_DeniedUserLoginNotRecorded keeps the write on the granted
+// path: a refused sign-in is not a login.
+func TestAuthCallback_DeniedUserLoginNotRecorded(t *testing.T) {
+	setup := setupAuthCallbackTest(t)
+	defer setup.cleanup()
+
+	ctx := context.Background()
+
+	setup.oidcServer.tokenSubject = "blockedUserId"
+	state := createTestState(t, setup.proxyService, "https://test-proxy.example.com/dashboard")
+
+	req := httptest.NewRequest(http.MethodGet, "/reverse-proxy/callback?code=test-auth-code&state="+url.QueryEscape(state), nil)
+	rec := httptest.NewRecorder()
+	setup.router.ServeHTTP(rec, req)
+	require.Equal(t, http.StatusFound, rec.Code)
+
+	after, err := setup.store.GetUserByUserID(ctx, store.LockingStrengthNone, "blockedUserId")
+	require.NoError(t, err)
+	require.Nil(t, after.LastLogin, "a denied user must not be recorded as having logged in")
+}
+
+func TestAuthCallback_UserDeniedByAccountStatus(t *testing.T) {
+	tests := []struct {
+		name            string
+		subject         string
+		expectErrorDesc string
+	}{
+		{
+			name:            "pending approval",
+			subject:         "pendingUserId",
+			expectErrorDesc: "Your account is pending approval by an administrator",
+		},
+		{
+			name:            "blocked",
+			subject:         "blockedUserId",
+			expectErrorDesc: "Your account is blocked",
+		},
+		{
+			name:            "unknown to management",
+			subject:         "userMissingFromStoreId",
+			expectErrorDesc: "Service configuration error",
+		},
+		{
+			// The account topology stays out of the browser-visible message.
+			name:            "belongs to another account",
+			subject:         "otherAccountUserId",
+			expectErrorDesc: "Service configuration error",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			setup := setupAuthCallbackTest(t)
+			defer setup.cleanup()
+
+			setup.oidcServer.tokenSubject = tt.subject
+
+			state := createTestState(t, setup.proxyService, "https://test-proxy.example.com/dashboard")
+
+			req := httptest.NewRequest(http.MethodGet, "/reverse-proxy/callback?code=test-auth-code&state="+url.QueryEscape(state), nil)
+			rec := httptest.NewRecorder()
+
+			setup.router.ServeHTTP(rec, req)
+
+			require.Equal(t, http.StatusFound, rec.Code)
+
+			parsedLocation, err := url.Parse(rec.Header().Get("Location"))
+			require.NoError(t, err)
+
+			require.Empty(t, parsedLocation.Query().Get("session_token"), "Denied user must not receive a session token")
+			require.Equal(t, "access_denied", parsedLocation.Query().Get("error"))
+			require.Equal(t, tt.expectErrorDesc, parsedLocation.Query().Get("error_description"))
+		})
+	}
+}
+
 func TestAuthCallback_ProxyNotFound(t *testing.T) {
 	setup := setupAuthCallbackTest(t)
 	defer setup.cleanup()

management/server/http/handlers/setup_keys/setupkeys_handler.go

@@ -60,6 +60,19 @@ func (h *handler) createSetupKey(w http.ResponseWriter, r *http.Request, userAut
 		return
 	}
 
+	// A one-off key can be used once, and GenerateSetupKey pins its usage limit
+	// at 1 whatever the request says. Silently overriding a caller that asked
+	// for a different number leaves them holding a key that does not do what
+	// they configured, and no way to find out except by using it. Only values
+	// above 1 are refused: usage_limit is a required field with no null, so 0
+	// cannot be told apart from a caller that has nothing to say about it.
+	if types.SetupKeyType(req.Type) == types.SetupKeyOneOff && req.UsageLimit > 1 {
+		util.WriteError(r.Context(), status.Errorf(status.InvalidArgument,
+			"usage_limit %d is not valid for a one-off setup key, which can be used once; use type reusable for a key that can be used more than once",
+			req.UsageLimit), w)
+		return
+	}
+
 	expiresIn := time.Duration(req.ExpiresIn) * time.Second
 
 	if expiresIn < 0 {

management/server/http/handlers/setup_keys/setupkeys_handler_test.go

@@ -135,6 +135,40 @@ func TestSetupKeysHandlers(t *testing.T) {
 			expectedBody:     true,
 			expectedSetupKey: expectedNewKey,
 		},
+		{
+			// A one-off key is used once. Asking for more used to be accepted
+			// and then quietly reduced to 1.
+			name:        "Create One-Off Setup Key With Conflicting Usage Limit",
+			requestType: http.MethodPost,
+			requestPath: "/api/setup-keys",
+			requestBody: bytes.NewBuffer(
+				[]byte(fmt.Sprintf("{\"name\":\"%s\",\"type\":\"one-off\",\"expires_in\":86400,\"usage_limit\":5}", newSetupKeyName))),
+			expectedStatus: http.StatusUnprocessableEntity,
+			expectedBody:   false,
+		},
+		{
+			// 0 is what a caller sends when it has nothing to say about the
+			// usage limit, since the field is required and has no null, so it
+			// has to keep working.
+			name:        "Create One-Off Setup Key Without Usage Limit",
+			requestType: http.MethodPost,
+			requestPath: "/api/setup-keys",
+			requestBody: bytes.NewBuffer(
+				[]byte(fmt.Sprintf("{\"name\":\"%s\",\"type\":\"one-off\",\"expires_in\":86400,\"usage_limit\":0}", newSetupKeyName))),
+			expectedStatus: http.StatusOK,
+			expectedBody:   false,
+		},
+		{
+			// Only one-off keys are constrained; a reusable key means what it
+			// says.
+			name:        "Create Reusable Setup Key With Usage Limit",
+			requestType: http.MethodPost,
+			requestPath: "/api/setup-keys",
+			requestBody: bytes.NewBuffer(
+				[]byte(fmt.Sprintf("{\"name\":\"%s\",\"type\":\"reusable\",\"expires_in\":86400,\"usage_limit\":5}", newSetupKeyName))),
+			expectedStatus: http.StatusOK,
+			expectedBody:   false,
+		},
 		{
 			name:        "Update Setup Key",
 			requestType: http.MethodPut,

management/server/http/handlers/users/users_handler.go

@@ -199,7 +199,7 @@ func (h *handler) getAllUsers(w http.ResponseWriter, r *http.Request, userAuth *
 		}
 
 		includeServiceUser, err := strconv.ParseBool(serviceUser)
-		log.WithContext(r.Context()).Debugf("Should include service user: %v", includeServiceUser)
+		log.WithContext(r.Context()).Tracef("Should include service user: %v", includeServiceUser)
 		if err != nil {
 			util.WriteError(r.Context(), status.Errorf(status.InvalidArgument, "invalid service_user query parameter"), w)
 			return

management/server/http/middleware/auth_middleware.go

@@ -11,8 +11,6 @@ import (
 	log "github.com/sirupsen/logrus"
 	"go.opentelemetry.io/otel/metric"
 
-	"github.com/netbirdio/management-integrations/integrations"
-
 	serverauth "github.com/netbirdio/netbird/management/server/auth"
 	nbcontext "github.com/netbirdio/netbird/management/server/context"
 	"github.com/netbirdio/netbird/management/server/http/middleware/bypass"
@@ -27,6 +25,8 @@ type SyncUserJWTGroupsFunc func(ctx context.Context, userAuth auth.UserAuth) err
 
 type GetUserFromUserAuthFunc func(ctx context.Context, userAuth auth.UserAuth) (*types.User, error)
 
+type IsValidChildAccountFunc func(ctx context.Context, userID, accountID, childAccountID string) bool
+
 // AuthMiddleware middleware to verify personal access tokens (PAT) and JWT tokens
 type AuthMiddleware struct {
 	authManager         serverauth.Manager
@@ -35,6 +35,7 @@ type AuthMiddleware struct {
 	syncUserJWTGroups   SyncUserJWTGroupsFunc
 	rateLimiter         *APIRateLimiter
 	patUsageTracker     *PATUsageTracker
+	isValidChildAccount IsValidChildAccountFunc
 }
 
 // NewAuthMiddleware instance constructor
@@ -45,6 +46,7 @@ func NewAuthMiddleware(
 	getUserFromUserAuth GetUserFromUserAuthFunc,
 	rateLimiter *APIRateLimiter,
 	meter metric.Meter,
+	isValidChildAccount IsValidChildAccountFunc,
 ) *AuthMiddleware {
 	var patUsageTracker *PATUsageTracker
 	if meter != nil {
@@ -62,6 +64,7 @@ func NewAuthMiddleware(
 		getUserFromUserAuth: getUserFromUserAuth,
 		rateLimiter:         rateLimiter,
 		patUsageTracker:     patUsageTracker,
+		isValidChildAccount: isValidChildAccount,
 	}
 }
 
@@ -124,7 +127,7 @@ func (m *AuthMiddleware) checkJWTFromRequest(r *http.Request, authHeaderParts []
 	}
 
 	if impersonate, ok := r.URL.Query()["account"]; ok && len(impersonate) == 1 {
-		if integrations.IsValidChildAccount(ctx, userAuth.UserId, userAuth.AccountId, impersonate[0]) {
+		if m.isValidChildAccount(ctx, userAuth.UserId, userAuth.AccountId, impersonate[0]) {
 			userAuth.AccountId = impersonate[0]
 			userAuth.IsChild = true
 		}
@@ -149,7 +152,11 @@ func (m *AuthMiddleware) checkJWTFromRequest(r *http.Request, authHeaderParts []
 		return err
 	}
 
-	err = m.syncUserJWTGroups(ctx, userAuth)
+	// Detach the group-sync write from the request's cancellation: the dashboard
+	// SPA aborts in-flight requests on re-render, which would otherwise cancel the
+	// DB transaction mid-write and silently drop the synced groups. Context values
+	// (request id, logger) are preserved; the store bounds the tx with its own timeout.
+	err = m.syncUserJWTGroups(context.WithoutCancel(ctx), userAuth)
 	if err != nil {
 		log.WithContext(ctx).Errorf("HTTP server failed to sync user JWT groups: %s", err)
 	}
@@ -203,7 +210,7 @@ func (m *AuthMiddleware) checkPATFromRequest(r *http.Request, authHeaderParts []
 	}
 
 	if impersonate, ok := r.URL.Query()["account"]; ok && len(impersonate) == 1 {
-		if integrations.IsValidChildAccount(r.Context(), userAuth.UserId, userAuth.AccountId, impersonate[0]) {
+		if m.isValidChildAccount(r.Context(), userAuth.UserId, userAuth.AccountId, impersonate[0]) {
 			userAuth.AccountId = impersonate[0]
 			userAuth.IsChild = true
 		}

management/server/http/middleware/auth_middleware_test.go

@@ -211,6 +211,7 @@ func TestAuthMiddleware_Handler(t *testing.T) {
 		},
 		disabledLimiter,
 		nil,
+		func(_ context.Context, _, _, _ string) bool { return false },
 	)
 
 	handlerToTest := authMiddleware.Handler(nextHandler)
@@ -240,6 +241,66 @@ func TestAuthMiddleware_Handler(t *testing.T) {
 	}
 }
 
+// TestAuthMiddleware_SyncUserJWTGroupsDetachedFromRequestCancellation ensures the
+// JWT group sync write is not bound to the request context. The dashboard SPA
+// routinely aborts in-flight requests on re-render/navigation; if the sync ran in
+// the request context, the cancellation would roll back the DB transaction and the
+// synced groups would silently never persist. The sync must receive a context that
+// is not cancelled even when the originating request is.
+func TestAuthMiddleware_SyncUserJWTGroupsDetachedFromRequestCancellation(t *testing.T) {
+	var (
+		syncCalled bool
+		syncCtxErr error
+	)
+
+	mockAuth := &auth.MockManager{
+		ValidateAndParseTokenFunc:       mockValidateAndParseToken,
+		EnsureUserAccessByJWTGroupsFunc: mockEnsureUserAccessByJWTGroups,
+		MarkPATUsedFunc:                 mockMarkPATUsed,
+		GetPATInfoFunc:                  mockGetAccountInfoFromPAT,
+	}
+
+	disabledLimiter := NewAPIRateLimiter(nil)
+	disabledLimiter.SetEnabled(false)
+
+	authMiddleware := NewAuthMiddleware(
+		mockAuth,
+		func(ctx context.Context, userAuth nbauth.UserAuth) (string, string, error) {
+			return userAuth.AccountId, userAuth.UserId, nil
+		},
+		func(ctx context.Context, userAuth nbauth.UserAuth) error {
+			syncCalled = true
+			syncCtxErr = ctx.Err()
+			return nil
+		},
+		func(ctx context.Context, userAuth nbauth.UserAuth) (*types.User, error) {
+			return &types.User{}, nil
+		},
+		disabledLimiter,
+		nil,
+		func(_ context.Context, _, _, _ string) bool { return false },
+	)
+
+	handlerToTest := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {}))
+
+	// Simulate the dashboard aborting the request: it arrives already cancelled.
+	ctx, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	req := httptest.NewRequest("GET", "http://testing/test", nil).WithContext(ctx)
+	req.Header.Set("Authorization", "Bearer "+JWT)
+	rec := httptest.NewRecorder()
+
+	handlerToTest.ServeHTTP(rec, req)
+
+	if !syncCalled {
+		t.Fatal("syncUserJWTGroups was not called")
+	}
+	if syncCtxErr != nil {
+		t.Fatalf("syncUserJWTGroups received a cancelled context (%v); the group-sync write must be detached from request cancellation", syncCtxErr)
+	}
+}
+
 func TestAuthMiddleware_RateLimiting(t *testing.T) {
 	mockAuth := &auth.MockManager{
 		ValidateAndParseTokenFunc:       mockValidateAndParseToken,
@@ -270,6 +331,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -322,6 +384,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -365,6 +428,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -409,6 +473,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -473,6 +538,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -532,6 +598,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -587,6 +654,7 @@ func TestAuthMiddleware_RateLimiting(t *testing.T) {
 			},
 			NewAPIRateLimiter(rateLimitConfig),
 			nil,
+			func(_ context.Context, _, _, _ string) bool { return false },
 		)
 
 		handler := authMiddleware.Handler(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -687,6 +755,7 @@ func TestAuthMiddleware_Handler_Child(t *testing.T) {
 		},
 		disabledLimiter,
 		nil,
+		func(_ context.Context, _, _, _ string) bool { return false },
 	)
 
 	for _, tc := range tt {

management/server/http/testing/integration/networks_handler_integration_test.go

@@ -1319,7 +1319,7 @@ func Test_NetworkRouters_Update(t *testing.T) {
 			},
 		},
 		{
-			name:      "Update non-existing router creates it",
+			name:      "Update non-existing router returns not found",
 			networkId: "testNetworkId",
 			routerId:  "nonExistingRouterId",
 			requestBody: &api.NetworkRouterRequest{
@@ -1328,11 +1328,7 @@ func Test_NetworkRouters_Update(t *testing.T) {
 				Metric:     100,
 				Enabled:    true,
 			},
-			expectedStatus: http.StatusOK,
-			verifyResponse: func(t *testing.T, router *api.NetworkRouter) {
-				t.Helper()
-				assert.Equal(t, "nonExistingRouterId", router.Id)
-			},
+			expectedStatus: http.StatusNotFound,
 		},
 		{
 			name:      "Update router with both peer and peer_groups",

management/server/http/testing/integration/setupkeys_handler_integration_test.go

@@ -136,7 +136,10 @@ func Test_SetupKeys_Create(t *testing.T) {
 			},
 		},
 		{
-			name:        "Create Setup Key as on-off with more than one usage",
+			// The key used to be created anyway, with its usage limit quietly
+			// reduced to 1, so the caller was told a key they had not asked for
+			// was what they asked for.
+			name:        "Create Setup Key as one-off with more than one usage",
 			requestType: http.MethodPost,
 			requestPath: "/api/setup-keys",
 			requestBody: &api.CreateSetupKeyRequest{
@@ -146,23 +149,7 @@ func Test_SetupKeys_Create(t *testing.T) {
 				Type:       "one-off",
 				UsageLimit: 3,
 			},
-			expectedStatus: http.StatusOK,
-			expectedResponse: &api.SetupKey{
-				AutoGroups: []string{},
-				Ephemeral:  false,
-				Expires:    time.Time{},
-				Id:         "",
-				Key:        "",
-				LastUsed:   time.Time{},
-				Name:       testing_tools.NewKeyName,
-				Revoked:    false,
-				State:      "valid",
-				Type:       "one-off",
-				UpdatedAt:  time.Now(),
-				UsageLimit: 1,
-				UsedTimes:  0,
-				Valid:      true,
-			},
+			expectedStatus: http.StatusUnprocessableEntity,
 		},
 		{
 			name:        "Create Setup Key with expiration in the past",

management/server/http/testing/testing_tools/channel/channel.go

@@ -7,6 +7,7 @@ import (
 	"time"
 
 	"github.com/golang-jwt/jwt/v5"
+	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 
 	"go.opentelemetry.io/otel/metric/noop"
@@ -57,13 +58,21 @@ func BuildApiBlackBoxWithDBState(t testing_tools.TB, sqlFile string, expectedPee
 	}
 	t.Cleanup(cleanup)
 
-	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
+	// Bound the background loops these managers start (account request buffer,
+	// telemetry P95 flushers, PAT usage tracker, API rate limiter, proxy service
+	// cleanup, cache janitors, DB connection pools) to the test's lifetime. On
+	// context.Background() they never stop and accumulate across the package,
+	// exhausting DB connections until the suite hits the 20m test timeout.
+	ctx, cancel := context.WithCancel(context.Background())
+	t.Cleanup(cancel)
+
+	metrics, err := telemetry.NewDefaultAppMetrics(ctx)
 	if err != nil {
 		t.Fatalf("Failed to create metrics: %v", err)
 	}
 
 	peersUpdateManager := update_channel.NewPeersUpdateManager(nil)
-	updMsg := peersUpdateManager.CreateChannel(context.Background(), testing_tools.TestPeerId)
+	updMsg := peersUpdateManager.CreateChannel(ctx, testing_tools.TestPeerId)
 	done := make(chan struct{})
 	if validateUpdate {
 		go func() {
@@ -86,15 +95,13 @@ func BuildApiBlackBoxWithDBState(t testing_tools.TB, sqlFile string, expectedPee
 
 	jobManager := job.NewJobManager(nil, store, peersManager)
 
-	ctx := context.Background()
-
 	cacheStore, err := nbcache.NewStore(ctx, 100*time.Millisecond, 300*time.Millisecond, 100)
 	if err != nil {
 		t.Fatalf("Failed to create cache store: %v", err)
 	}
 
 	requestBuffer := server.NewAccountRequestBuffer(ctx, store)
-	networkMapController := controller.NewController(ctx, store, metrics, peersUpdateManager, requestBuffer, server.MockIntegratedValidator{}, settingsManager, "", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peersManager), &config.Config{})
+	networkMapController := controller.NewController(ctx, store, metrics, peersUpdateManager, requestBuffer, server.MockIntegratedValidator{}, settingsManager, "", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peersManager), &config.Config{}, nil)
 	am, err := server.BuildManager(ctx, nil, store, networkMapController, jobManager, nil, "", &activity.InMemoryEventStore{}, geoMock, false, validatorMock, metrics, proxyController, settingsManager, permissionsManager, false, cacheStore)
 	if err != nil {
 		t.Fatalf("Failed to create manager: %v", err)
@@ -108,7 +115,11 @@ func BuildApiBlackBoxWithDBState(t testing_tools.TB, sqlFile string, expectedPee
 	if err != nil {
 		t.Fatalf("Failed to create proxy manager: %v", err)
 	}
-	proxyServiceServer := nbgrpc.NewProxyServiceServer(accessLogsManager, proxyTokenStore, pkceverifierStore, nbgrpc.ProxyOIDCConfig{}, peersManager, userManager, proxyMgr)
+	proxyServiceServer := nbgrpc.NewProxyServiceServer(accessLogsManager, proxyTokenStore, pkceverifierStore, nbgrpc.ProxyOIDCConfig{}, peersManager, userManager, nil, proxyMgr, nil)
+	// NewProxyServiceServer starts cleanupStaleProxies on a context it derives
+	// from context.Background(), independent of the cancellable ctx above;
+	// Close() cancels it so the goroutine does not outlive the test.
+	t.Cleanup(proxyServiceServer.Close)
 	domainManager := manager.NewManager(store, proxyMgr, am)
 	serviceProxyController, err := proxymanager.NewGRPCController(proxyServiceServer, noopMeter)
 	if err != nil {
@@ -134,7 +145,8 @@ func BuildApiBlackBoxWithDBState(t testing_tools.TB, sqlFile string, expectedPee
 	customZonesManager := zonesManager.NewManager(store, am, "")
 	zoneRecordsManager := recordsManager.NewManager(store, am)
 
-	apiHandler, err := http2.NewAPIHandler(context.Background(), am, networksManager, resourcesManager, routersManager, groupsManager, geoMock, authManagerMock, metrics, validatorMock, proxyController, permissionsManager, peersManager, settingsManager, customZonesManager, zoneRecordsManager, networkMapController, nil, serviceManager, &domainManager, accessLogsManager, nil, nil, nil)
+	apiRouter := mux.NewRouter().PathPrefix("/api").Subrouter()
+	apiHandler, err := http2.NewAPIHandler(ctx, apiRouter, am, networksManager, resourcesManager, routersManager, groupsManager, geoMock, authManagerMock, metrics, permissionsManager, settingsManager, customZonesManager, zoneRecordsManager, networkMapController, nil, serviceManager, &domainManager, accessLogsManager, nil, nil, nil, nil, nil)
 	if err != nil {
 		t.Fatalf("Failed to create API handler: %v", err)
 	}
@@ -197,13 +209,21 @@ func BuildApiBlackBoxWithDBStateAndPeerChannel(t testing_tools.TB, sqlFile strin
 	}
 	t.Cleanup(cleanup)
 
-	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
+	// Bound the background loops these managers start (account request buffer,
+	// telemetry P95 flushers, PAT usage tracker, API rate limiter, proxy service
+	// cleanup, cache janitors, DB connection pools) to the test's lifetime. On
+	// context.Background() they never stop and accumulate across the package,
+	// exhausting DB connections until the suite hits the 20m test timeout.
+	ctx, cancel := context.WithCancel(context.Background())
+	t.Cleanup(cancel)
+
+	metrics, err := telemetry.NewDefaultAppMetrics(ctx)
 	if err != nil {
 		t.Fatalf("Failed to create metrics: %v", err)
 	}
 
 	peersUpdateManager := update_channel.NewPeersUpdateManager(nil)
-	updMsg := peersUpdateManager.CreateChannel(context.Background(), testing_tools.TestPeerId)
+	updMsg := peersUpdateManager.CreateChannel(ctx, testing_tools.TestPeerId)
 
 	geoMock := &geolocation.Mock{}
 	validatorMock := server.MockIntegratedValidator{}
@@ -215,15 +235,13 @@ func BuildApiBlackBoxWithDBStateAndPeerChannel(t testing_tools.TB, sqlFile strin
 
 	jobManager := job.NewJobManager(nil, store, peersManager)
 
-	ctx := context.Background()
-
 	cacheStore, err := nbcache.NewStore(ctx, 100*time.Millisecond, 300*time.Millisecond, 100)
 	if err != nil {
 		t.Fatalf("Failed to create cache store: %v", err)
 	}
 
 	requestBuffer := server.NewAccountRequestBuffer(ctx, store)
-	networkMapController := controller.NewController(ctx, store, metrics, peersUpdateManager, requestBuffer, server.MockIntegratedValidator{}, settingsManager, "", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peersManager), &config.Config{})
+	networkMapController := controller.NewController(ctx, store, metrics, peersUpdateManager, requestBuffer, server.MockIntegratedValidator{}, settingsManager, "", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peersManager), &config.Config{}, nil)
 	am, err := server.BuildManager(ctx, nil, store, networkMapController, jobManager, nil, "", &activity.InMemoryEventStore{}, geoMock, false, validatorMock, metrics, proxyController, settingsManager, permissionsManager, false, cacheStore)
 	if err != nil {
 		t.Fatalf("Failed to create manager: %v", err)
@@ -237,7 +255,11 @@ func BuildApiBlackBoxWithDBStateAndPeerChannel(t testing_tools.TB, sqlFile strin
 	if err != nil {
 		t.Fatalf("Failed to create proxy manager: %v", err)
 	}
-	proxyServiceServer := nbgrpc.NewProxyServiceServer(accessLogsManager, proxyTokenStore, pkceverifierStore, nbgrpc.ProxyOIDCConfig{}, peersManager, userManager, proxyMgr)
+	proxyServiceServer := nbgrpc.NewProxyServiceServer(accessLogsManager, proxyTokenStore, pkceverifierStore, nbgrpc.ProxyOIDCConfig{}, peersManager, userManager, nil, proxyMgr, nil)
+	// NewProxyServiceServer starts cleanupStaleProxies on a context it derives
+	// from context.Background(), independent of the cancellable ctx above;
+	// Close() cancels it so the goroutine does not outlive the test.
+	t.Cleanup(proxyServiceServer.Close)
 	domainManager := manager.NewManager(store, proxyMgr, am)
 	serviceProxyController, err := proxymanager.NewGRPCController(proxyServiceServer, noopMeter)
 	if err != nil {
@@ -263,7 +285,8 @@ func BuildApiBlackBoxWithDBStateAndPeerChannel(t testing_tools.TB, sqlFile strin
 	customZonesManager := zonesManager.NewManager(store, am, "")
 	zoneRecordsManager := recordsManager.NewManager(store, am)
 
-	apiHandler, err := http2.NewAPIHandler(context.Background(), am, networksManager, resourcesManager, routersManager, groupsManager, geoMock, authManagerMock, metrics, validatorMock, proxyController, permissionsManager, peersManager, settingsManager, customZonesManager, zoneRecordsManager, networkMapController, nil, serviceManager, &domainManager, accessLogsManager, nil, nil, nil)
+	apiRouter := mux.NewRouter().PathPrefix("/api").Subrouter()
+	apiHandler, err := http2.NewAPIHandler(ctx, apiRouter, am, networksManager, resourcesManager, routersManager, groupsManager, geoMock, authManagerMock, metrics, permissionsManager, settingsManager, customZonesManager, zoneRecordsManager, networkMapController, nil, serviceManager, &domainManager, accessLogsManager, nil, nil, nil, nil, nil)
 	if err != nil {
 		t.Fatalf("Failed to create API handler: %v", err)
 	}

management/server/http/testing/testing_tools/tools.go

@@ -5,9 +5,9 @@ import (
 	"context"
 	"fmt"
 	"io"
-	"net"
 	"net/http"
 	"net/http/httptest"
+	"net/netip"
 	"os"
 	"strconv"
 	"testing"
@@ -137,7 +137,7 @@ func PopulateTestData(b *testing.B, am account.Manager, peers, groups, users, se
 			ID:       fmt.Sprintf("oldpeer-%d", i),
 			DNSLabel: fmt.Sprintf("oldpeer-%d", i),
 			Key:      peerKey.PublicKey().String(),
-			IP:       net.ParseIP(fmt.Sprintf("100.64.%d.%d", i/256, i%256)),
+			IP:       netip.MustParseAddr(fmt.Sprintf("100.64.%d.%d", i/256, i%256)),
 			Status:   &nbpeer.PeerStatus{LastSeen: time.Now().UTC(), Connected: true},
 			UserID:   TestUserId,
 		}

management/server/identity_provider.go

@@ -232,7 +232,7 @@ func identityProviderToConnectorConfig(idpConfig *types.IdentityProvider) *dex.C
 }
 
 // generateIdentityProviderID generates a unique ID for an identity provider.
-// For specific provider types (okta, zitadel, entra, google, pocketid, microsoft),
+// For specific provider types (okta, zitadel, entra, google, pocketid, microsoft, adfs),
 // the ID is prefixed with the type name. Generic OIDC providers get no prefix.
 func generateIdentityProviderID(idpType types.IdentityProviderType) string {
 	id := xid.New().String()
@@ -254,6 +254,8 @@ func generateIdentityProviderID(idpType types.IdentityProviderType) string {
 		return "authentik-" + id
 	case types.IdentityProviderTypeKeycloak:
 		return "keycloak-" + id
+	case types.IdentityProviderTypeADFS:
+		return "adfs-" + id
 	default:
 		// Generic OIDC - no prefix
 		return id

management/server/identity_provider_test.go

@@ -10,9 +10,9 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+	"go.uber.org/mock/gomock"
 
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
@@ -34,6 +34,20 @@ import (
 
 func createManagerWithEmbeddedIdP(t testing.TB) (*DefaultAccountManager, *update_channel.PeersUpdateManager, error) {
 	t.Helper()
+	return createManagerWithEmbeddedIdPMode(t, "netbird.selfhosted")
+}
+
+func createManagerWithEmbeddedIdPMode(t testing.TB, singleAccountModeDomain string) (*DefaultAccountManager, *update_channel.PeersUpdateManager, error) {
+	t.Helper()
+	return createManagerWithEmbeddedIdPModeAndSetup(t, singleAccountModeDomain, nil)
+}
+
+func createManagerWithEmbeddedIdPModeAndSetup(
+	t testing.TB,
+	singleAccountModeDomain string,
+	setupStore func(context.Context, store.Store) error,
+) (*DefaultAccountManager, *update_channel.PeersUpdateManager, error) {
+	t.Helper()
 
 	ctx := context.Background()
 
@@ -43,6 +57,11 @@ func createManagerWithEmbeddedIdP(t testing.TB) (*DefaultAccountManager, *update
 		return nil, nil, err
 	}
 	t.Cleanup(cleanUp)
+	if setupStore != nil {
+		if err := setupStore(ctx, testStore); err != nil {
+			return nil, nil, err
+		}
+	}
 
 	// Create embedded IdP manager
 	embeddedConfig := &idp.EmbeddedIdPConfig{
@@ -92,8 +111,8 @@ func createManagerWithEmbeddedIdP(t testing.TB) (*DefaultAccountManager, *update
 
 	updateManager := update_channel.NewPeersUpdateManager(metrics)
 	requestBuffer := NewAccountRequestBuffer(ctx, testStore)
-	networkMapController := controller.NewController(ctx, testStore, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.cloud", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(testStore, peersManager), &config.Config{})
-	manager, err := BuildManager(ctx, &config.Config{}, testStore, networkMapController, job.NewJobManager(nil, testStore, peersManager), idpManager, "", eventStore, nil, false, MockIntegratedValidator{}, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
+	networkMapController := controller.NewController(ctx, testStore, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.cloud", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(testStore, peersManager), &config.Config{}, nil)
+	manager, err := BuildManager(ctx, &config.Config{}, testStore, networkMapController, job.NewJobManager(nil, testStore, peersManager), idpManager, singleAccountModeDomain, eventStore, nil, false, MockIntegratedValidator{}, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -196,6 +215,23 @@ func TestDefaultAccountManager_GetIdentityProvider_NotFound(t *testing.T) {
 	assert.Contains(t, err.Error(), "not found")
 }
 
+func TestUpdateUserAuthWithSingleModeKeepsConfiguredDomain(t *testing.T) {
+	ctx := context.Background()
+	manager, _, err := createManagerWithEmbeddedIdPModeAndSetup(t, "netbird.selfhosted", func(ctx context.Context, testStore store.Store) error {
+		// An account with no domain, as left behind by an IdP that emitted no domain claims.
+		return testStore.SaveAccount(ctx, newAccountWithId(ctx, "account-1", "user-1", "", "", "", false))
+	})
+	require.NoError(t, err)
+	require.True(t, manager.singleAccountMode)
+
+	userAuth := auth.UserAuth{UserId: "user-2"}
+	require.NoError(t, manager.updateUserAuthWithSingleMode(ctx, &userAuth))
+
+	assert.Equal(t, "netbird.selfhosted", userAuth.Domain,
+		"An empty account domain must not clear the configured single account domain")
+	assert.Equal(t, types.PrivateCategory, userAuth.DomainCategory)
+}
+
 func TestDefaultAccountManager_UpdateIdentityProvider_Validation(t *testing.T) {
 	manager, _, err := createManager(t)
 	require.NoError(t, err)

management/server/idp/embedded.go

@@ -2,9 +2,13 @@ package idp
 
 import (
 	"context"
+	"encoding/base64"
 	"errors"
 	"fmt"
 	"net/http"
+	"net/url"
+	"os"
+	"path"
 	"strings"
 
 	"github.com/dexidp/dex/storage"
@@ -17,12 +21,16 @@ import (
 )
 
 const (
-	staticClientDashboard  = "netbird-dashboard"
-	staticClientCLI        = "netbird-cli"
-	defaultCLIRedirectURL1 = "http://localhost:53000/"
-	defaultCLIRedirectURL2 = "http://localhost:54000/"
-	defaultScopes          = "openid profile email groups"
-	defaultUserIDClaim     = "sub"
+	StaticClientDashboard      = "netbird-dashboard"
+	StaticClientCLI            = "netbird-cli"
+	DefaultTOTPAuthenticatorID = "default-totp"
+	LocalConnectorID           = dex.LocalConnectorID
+
+	defaultCLIRedirectURL1        = "http://localhost:53000/"
+	defaultCLIRedirectURL2        = "http://localhost:54000/"
+	defaultScopes                 = "openid profile email groups"
+	defaultUserIDClaim            = "sub"
+	sessionCookieEncryptionKeyEnv = "NB_IDP_SESSION_COOKIE_ENCRYPTION_KEY"
 )
 
 // EmbeddedIdPConfig contains configuration for the embedded Dex OIDC identity provider
@@ -49,8 +57,31 @@ type EmbeddedIdPConfig struct {
 	// Existing local users are preserved and will be able to login again if re-enabled.
 	// Cannot be enabled if no external identity provider connectors are configured.
 	LocalAuthDisabled bool
+	// MfaSessionMaxLifetime is the maximum MFA session duration from creation (e.g. "24h").
+	// Defaults to "24h" if empty.
+	MfaSessionMaxLifetime string
+	// MfaSessionIdleTimeout is the idle timeout after which the MFA session expires (e.g. "1h").
+	// Defaults to "1h" if empty.
+	MfaSessionIdleTimeout string
+	// MfaSessionRememberMe controls the default state of the "remember me" checkbox on the
+	// login screen. When true, the session cookie persists across browser tabs/restarts so
+	// MFA is not re-prompted until the session expires. Defaults to false.
+	MfaSessionRememberMe bool
+	// SessionCookieEncryptionKey is the optional AES key used to encrypt embedded IdP session cookies.
+	// It can also be set with NB_IDP_SESSION_COOKIE_ENCRYPTION_KEY. The value must be 16, 24, or 32
+	// bytes when provided as a raw string, or base64-encoded to one of those lengths.
+	SessionCookieEncryptionKey string
+	// Dashboard Post logout redirect URIs, these are required to tell
+	// Dex what to allow when an RP-Initiated logout is started by the frontend
+	// at least one of these must match the dashboard base URL or the dashboard
+	// DASHBOARD_POST_LOGOUT_URL environment variable
+	// WARNING: Dex only uses exact match, not wildcards
+	DashboardPostLogoutRedirectURIs []string
 	// StaticConnectors are additional connectors to seed during initialization
 	StaticConnectors []dex.Connector
+	// GrantTypes restricts allowed OAuth2 grants; empty means all (Dex default). Omit the
+	// device_code grant to disable the device flow; keep authorization_code and refresh_token.
+	GrantTypes []string
 }
 
 // EmbeddedStorageConfig holds storage configuration for the embedded IdP.
@@ -115,10 +146,13 @@ func (c *EmbeddedIdPConfig) ToYAMLConfig() (*dex.YAMLConfig, error) {
 		return nil, fmt.Errorf("invalid IdP storage config: %w", err)
 	}
 
-	// Build CLI redirect URIs including the device callback (both relative and absolute)
+	// Build CLI redirect URIs including the device callback. Dex uses the issuer-relative
+	// path (for example, /oauth2/device/callback) when completing the device flow, so
+	// include it explicitly in addition to the legacy bare path and absolute URL.
 	cliRedirectURIs := c.CLIRedirectURIs
 	cliRedirectURIs = append(cliRedirectURIs, "/device/callback")
-	cliRedirectURIs = append(cliRedirectURIs, c.Issuer+"/device/callback")
+	cliRedirectURIs = append(cliRedirectURIs, issuerRelativeDeviceCallback(c.Issuer))
+	cliRedirectURIs = append(cliRedirectURIs, strings.TrimSuffix(c.Issuer, "/")+"/device/callback")
 
 	// Build dashboard redirect URIs including the OAuth callback for proxy authentication
 	dashboardRedirectURIs := c.DashboardRedirectURIs
@@ -126,6 +160,15 @@ func (c *EmbeddedIdPConfig) ToYAMLConfig() (*dex.YAMLConfig, error) {
 	// todo: resolve import cycle
 	dashboardRedirectURIs = append(dashboardRedirectURIs, baseURL+"/api/reverse-proxy/callback")
 
+	dashboardPostLogoutRedirectURIs := c.DashboardPostLogoutRedirectURIs
+	// It is safe to assume that most installations will share the location of the
+	// MGMT api and the dashboard, adding baseURL means less configuration for the instance admin
+	dashboardPostLogoutRedirectURIs = append(dashboardPostLogoutRedirectURIs, baseURL)
+
+	redirectURIs := make([]string, 0)
+	redirectURIs = append(redirectURIs, cliRedirectURIs...)
+	redirectURIs = append(redirectURIs, dashboardRedirectURIs...)
+
 	cfg := &dex.YAMLConfig{
 		Issuer: c.Issuer,
 		Storage: dex.Storage{
@@ -138,6 +181,7 @@ func (c *EmbeddedIdPConfig) ToYAMLConfig() (*dex.YAMLConfig, error) {
 		},
 		OAuth2: dex.OAuth2{
 			SkipApprovalScreen: true,
+			GrantTypes:         c.GrantTypes,
 		},
 		Frontend: dex.Frontend{
 			Issuer: "NetBird",
@@ -148,21 +192,28 @@ func (c *EmbeddedIdPConfig) ToYAMLConfig() (*dex.YAMLConfig, error) {
 		EnablePasswordDB: true,
 		StaticClients: []storage.Client{
 			{
-				ID:           staticClientDashboard,
-				Name:         "NetBird Dashboard",
-				Public:       true,
-				RedirectURIs: dashboardRedirectURIs,
+				ID:                     StaticClientDashboard,
+				Name:                   "NetBird Dashboard",
+				Public:                 true,
+				RedirectURIs:           redirectURIs,
+				PostLogoutRedirectURIs: sanitizePostLogoutRedirectURIs(dashboardPostLogoutRedirectURIs),
 			},
 			{
-				ID:           staticClientCLI,
+				ID:           StaticClientCLI,
 				Name:         "NetBird CLI",
 				Public:       true,
-				RedirectURIs: cliRedirectURIs,
+				RedirectURIs: redirectURIs,
 			},
 		},
 		StaticConnectors: c.StaticConnectors,
 	}
 
+	// Always initialize MFA providers and sessions so TOTP can be toggled at runtime.
+	// MFAChain on clients is NOT set here — it's synced from the DB setting on startup.
+	if err := configureMFA(cfg, c.MfaSessionMaxLifetime, c.MfaSessionIdleTimeout, c.MfaSessionRememberMe, c.SessionCookieEncryptionKey); err != nil {
+		return nil, err
+	}
+
 	// Add owner user if provided
 	if c.Owner != nil && c.Owner.Email != "" && c.Owner.Hash != "" {
 		username := c.Owner.Username
@@ -182,6 +233,108 @@ func (c *EmbeddedIdPConfig) ToYAMLConfig() (*dex.YAMLConfig, error) {
 	return cfg, nil
 }
 
+func issuerRelativeDeviceCallback(issuer string) string {
+	u, err := url.Parse(issuer)
+	if err != nil || u.Path == "" {
+		return "/device/callback"
+	}
+	return path.Join(u.Path, "/device/callback")
+}
+
+// Due to how the frontend generates the logout, sometimes it appends a trailing slash
+// and because Dex only allows exact matches, we need to make sure we always have both
+// versions of each provided uri
+func sanitizePostLogoutRedirectURIs(uris []string) []string {
+	result := make([]string, 0)
+	for _, uri := range uris {
+		if strings.HasSuffix(uri, "/") {
+			result = append(result, uri)
+			result = append(result, strings.TrimSuffix(uri, "/"))
+		} else {
+			result = append(result, uri)
+			result = append(result, uri+"/")
+		}
+	}
+
+	return result
+}
+
+func configureMFA(cfg *dex.YAMLConfig, sessionMaxLifetime, sessionIdleTimeout string, rememberMe bool, sessionCookieEncryptionKey string) error {
+	cfg.MFA.Authenticators = []dex.MFAAuthenticator{{
+		ID: DefaultTOTPAuthenticatorID,
+		// Has to be caps otherwise it will fail
+		Type: "TOTP",
+		Config: map[string]interface{}{
+			"issuer": "NetBird",
+		},
+		ConnectorTypes: []string{LocalConnectorID},
+	}}
+
+	if sessionMaxLifetime == "" {
+		sessionMaxLifetime = "24h"
+	}
+	if sessionIdleTimeout == "" {
+		sessionIdleTimeout = "1h"
+	}
+
+	cookieEncryptionKey, err := resolveSessionCookieEncryptionKey(sessionCookieEncryptionKey)
+	if err != nil {
+		return err
+	}
+
+	cfg.Sessions = &dex.Sessions{
+		CookieName:                 "netbird-session",
+		AbsoluteLifetime:           sessionMaxLifetime,
+		ValidIfNotUsedFor:          sessionIdleTimeout,
+		RememberMeCheckedByDefault: &rememberMe,
+		SSOSharedWithDefault:       "all",
+		CookieEncryptionKey:        cookieEncryptionKey,
+	}
+	// Absolutely required, otherwise the dex server will omit the MFA configuration entirely
+	os.Setenv("DEX_SESSIONS_ENABLED", "true")
+
+	// Note: MFAChain on clients is NOT set here.
+	// It is toggled at runtime via SetMFAEnabled() based on the account settings DB value.
+	return nil
+}
+
+func resolveSessionCookieEncryptionKey(configuredKey string) (string, error) {
+	key := strings.TrimSpace(configuredKey)
+	if key == "" {
+		key = strings.TrimSpace(os.Getenv(sessionCookieEncryptionKeyEnv))
+	}
+	if key == "" {
+		return "", nil
+	}
+
+	if validSessionCookieEncryptionKeyLength(len([]byte(key))) {
+		return key, nil
+	}
+
+	for _, encoding := range []*base64.Encoding{
+		base64.StdEncoding,
+		base64.RawStdEncoding,
+		base64.URLEncoding,
+		base64.RawURLEncoding,
+	} {
+		decoded, err := encoding.DecodeString(key)
+		if err == nil && validSessionCookieEncryptionKeyLength(len(decoded)) {
+			return string(decoded), nil
+		}
+	}
+
+	return "", fmt.Errorf("invalid embedded IdP session cookie encryption key:%s (or sessionCookieEncryptionKey) must be 16, 24, or 32 bytes as a raw string or base64-encoded to one of those lengths; got %d raw bytes", sessionCookieEncryptionKeyEnv, len([]byte(key)))
+}
+
+func validSessionCookieEncryptionKeyLength(length int) bool {
+	switch length {
+	case 16, 24, 32:
+		return true
+	default:
+		return false
+	}
+}
+
 // Compile-time check that EmbeddedIdPManager implements Manager interface
 var _ Manager = (*EmbeddedIdPManager)(nil)
 
@@ -215,6 +368,7 @@ type EmbeddedIdPManager struct {
 	provider   *dex.Provider
 	appMetrics telemetry.AppMetrics
 	config     EmbeddedIdPConfig
+	mfaEnabled bool
 }
 
 // NewEmbeddedIdPManager creates a new instance of EmbeddedIdPManager from a configuration.
@@ -589,7 +743,7 @@ func (m *EmbeddedIdPManager) GetDefaultScopes() string {
 
 // GetCLIClientID returns the client ID for CLI authentication.
 func (m *EmbeddedIdPManager) GetCLIClientID() string {
-	return staticClientCLI
+	return StaticClientCLI
 }
 
 // GetCLIRedirectURLs returns the redirect URLs configured for the CLI client.
@@ -628,7 +782,7 @@ func (m *EmbeddedIdPManager) GetLocalKeysLocation() string {
 
 // GetClientIDs returns the OAuth2 client IDs configured for this provider.
 func (m *EmbeddedIdPManager) GetClientIDs() []string {
-	return []string{staticClientDashboard, staticClientCLI}
+	return []string{StaticClientDashboard, StaticClientCLI}
 }
 
 // GetUserIDClaim returns the JWT claim name used for user identification.
@@ -641,6 +795,27 @@ func (m *EmbeddedIdPManager) IsLocalAuthDisabled() bool {
 	return m.config.LocalAuthDisabled
 }
 
+// SetMFAEnabled enables or disables TOTP MFA for local users by updating the MFAChain on OAuth2 clients.
+func (m *EmbeddedIdPManager) SetMFAEnabled(ctx context.Context, enabled bool) error {
+	var mfaChain []string
+	if enabled {
+		mfaChain = []string{DefaultTOTPAuthenticatorID}
+	}
+	if err := m.provider.SetClientsMFAChain(ctx, []string{
+		StaticClientCLI,
+		StaticClientDashboard,
+	}, mfaChain); err != nil {
+		return fmt.Errorf("failed to set MFA enabled=%v: %w", enabled, err)
+	}
+	m.mfaEnabled = enabled
+	return nil
+}
+
+// IsMFAEnabled returns whether TOTP MFA is currently enabled for local users.
+func (m *EmbeddedIdPManager) IsMFAEnabled() bool {
+	return m.mfaEnabled
+}
+
 // HasNonLocalConnectors checks if there are any identity provider connectors other than local.
 func (m *EmbeddedIdPManager) HasNonLocalConnectors(ctx context.Context) (bool, error) {
 	return m.provider.HasNonLocalConnectors(ctx)

management/server/idp/embedded_test.go

@@ -2,6 +2,7 @@ package idp
 
 import (
 	"context"
+	"encoding/base64"
 	"os"
 	"path/filepath"
 	"testing"
@@ -313,6 +314,100 @@ func TestEmbeddedIdPManager_UpdateUserPassword(t *testing.T) {
 	})
 }
 
+func TestEmbeddedIdPConfig_ToYAMLConfig_IncludesDeviceCallbackRedirectURI(t *testing.T) {
+	config := &EmbeddedIdPConfig{
+		Enabled: true,
+		Issuer:  "https://example.com/oauth2",
+		Storage: EmbeddedStorageConfig{
+			Type: "sqlite3",
+			Config: EmbeddedStorageTypeConfig{
+				File: filepath.Join(t.TempDir(), "dex.db"),
+			},
+		},
+	}
+
+	yamlConfig, err := config.ToYAMLConfig()
+	require.NoError(t, err)
+
+	var cliRedirectURIs []string
+	for _, client := range yamlConfig.StaticClients {
+		if client.ID == StaticClientCLI {
+			cliRedirectURIs = client.RedirectURIs
+			break
+		}
+	}
+	require.NotEmpty(t, cliRedirectURIs)
+	assert.Contains(t, cliRedirectURIs, "/device/callback")
+	assert.Contains(t, cliRedirectURIs, "/oauth2/device/callback")
+	assert.Contains(t, cliRedirectURIs, "https://example.com/oauth2/device/callback")
+}
+
+func TestEmbeddedIdPConfig_ToYAMLConfig_SessionCookieEncryptionKey(t *testing.T) {
+	t.Setenv(sessionCookieEncryptionKeyEnv, "")
+
+	rawKey := "0123456789abcdef0123456789abcdef"
+	config := &EmbeddedIdPConfig{
+		Enabled:                    true,
+		Issuer:                     "http://localhost:5556/dex",
+		SessionCookieEncryptionKey: base64.StdEncoding.EncodeToString([]byte(rawKey)),
+		Storage: EmbeddedStorageConfig{
+			Type: "sqlite3",
+			Config: EmbeddedStorageTypeConfig{
+				File: filepath.Join(t.TempDir(), "dex.db"),
+			},
+		},
+	}
+
+	yamlConfig, err := config.ToYAMLConfig()
+	require.NoError(t, err)
+	require.NotNil(t, yamlConfig.Sessions)
+	assert.Equal(t, rawKey, yamlConfig.Sessions.CookieEncryptionKey)
+}
+
+func TestResolveSessionCookieEncryptionKey(t *testing.T) {
+	rawKey := "0123456789abcdef0123456789abcdef"
+
+	t.Run("uses raw configured key", func(t *testing.T) {
+		t.Setenv(sessionCookieEncryptionKeyEnv, "")
+
+		key, err := resolveSessionCookieEncryptionKey(rawKey)
+		require.NoError(t, err)
+		assert.Equal(t, rawKey, key)
+	})
+
+	t.Run("uses base64 configured key", func(t *testing.T) {
+		t.Setenv(sessionCookieEncryptionKeyEnv, "")
+
+		key, err := resolveSessionCookieEncryptionKey(base64.StdEncoding.EncodeToString([]byte(rawKey)))
+		require.NoError(t, err)
+		assert.Equal(t, rawKey, key)
+	})
+
+	t.Run("falls back to env var", func(t *testing.T) {
+		t.Setenv(sessionCookieEncryptionKeyEnv, rawKey)
+
+		key, err := resolveSessionCookieEncryptionKey("")
+		require.NoError(t, err)
+		assert.Equal(t, rawKey, key)
+	})
+
+	t.Run("empty key disables encryption", func(t *testing.T) {
+		t.Setenv(sessionCookieEncryptionKeyEnv, "")
+
+		key, err := resolveSessionCookieEncryptionKey("")
+		require.NoError(t, err)
+		assert.Empty(t, key)
+	})
+
+	t.Run("rejects invalid key length", func(t *testing.T) {
+		t.Setenv(sessionCookieEncryptionKeyEnv, "")
+
+		_, err := resolveSessionCookieEncryptionKey("32")
+		require.Error(t, err)
+		assert.Contains(t, err.Error(), sessionCookieEncryptionKeyEnv)
+	})
+}
+
 func TestEmbeddedIdPManager_GetLocalKeysLocation(t *testing.T) {
 	ctx := context.Background()
 

management/server/idp/migration/migration.go

@@ -10,6 +10,8 @@ import (
 	"errors"
 	"fmt"
 	"os"
+	"regexp"
+	"strings"
 
 	log "github.com/sirupsen/logrus"
 
@@ -25,8 +27,10 @@ type Server interface {
 	EventStore() EventStore // may return nil
 }
 
-const idpSeedInfoKey = "IDP_SEED_INFO"
-const dryRunEnvKey = "NB_IDP_MIGRATION_DRY_RUN"
+const (
+	idpSeedInfoKey = "IDP_SEED_INFO"
+	dryRunEnvKey   = "NB_IDP_MIGRATION_DRY_RUN"
+)
 
 func isDryRun() bool {
 	return os.Getenv(dryRunEnvKey) == "true"
@@ -233,3 +237,163 @@ func PopulateUserInfo(s Server, idpManager idp.Manager, dryRun bool) error {
 
 	return nil
 }
+
+const DefaultSingleAccountDomain = "netbird.selfhosted"
+
+var (
+	ErrMultipleAccounts = errors.New("the embedded IdP supports a single account only")
+	ErrUnusableDomain   = errors.New("domain cannot be resolved in single account mode")
+	ErrDomainConflict   = errors.New("requested domain conflicts with the account domain")
+)
+
+var resolvableDomainRegexp = regexp.MustCompile(`^([a-z0-9]+(-[a-z0-9]+)*\.)+[a-z]{2,}$`)
+
+// RequireSingleAccount refuses to migrate an instance that holds more than one account.
+func RequireSingleAccount(s Server) error {
+	accountsCounter, err := s.Store().GetAccountsCounter(context.Background())
+	if err != nil {
+		return fmt.Errorf("failed to count accounts: %w", err)
+	}
+
+	if accountsCounter > 1 {
+		return errMultipleAccounts(accountsCounter)
+	}
+
+	return nil
+}
+
+func errMultipleAccounts(accountsCounter int64) error {
+	return fmt.Errorf("%w: this instance has %d accounts. Identity provider connectors are stored without "+
+		"an account scope, so every account would share and be able to manage the same connectors. "+
+		"Consolidate this instance to a single account, or keep using an external IdP, before migrating",
+		ErrMultipleAccounts, accountsCounter)
+}
+
+func NormalizeSingleAccountDomain(singleAccountDomain string) (string, error) {
+	if singleAccountDomain == "" {
+		singleAccountDomain = DefaultSingleAccountDomain
+	}
+
+	singleAccountDomain = strings.ToLower(singleAccountDomain)
+	if !resolvableDomainRegexp.MatchString(singleAccountDomain) {
+		return "", fmt.Errorf("%w: %q must contain at least one dot and only lowercase letters, digits and "+
+			"hyphens, otherwise users cannot join the existing account", ErrUnusableDomain, singleAccountDomain)
+	}
+
+	return singleAccountDomain, nil
+}
+
+// resolveAccountDomain picks the domain the account should end up with. The account keeps a usable
+// domain of its own, the configured one only fills a blank. Anything else is a conflict to report.
+func resolveAccountDomain(accountID, accountDomain, singleAccountDomain string, requested bool) (string, error) {
+	accountDomain = strings.ToLower(accountDomain)
+
+	if accountDomain == "" {
+		return singleAccountDomain, nil
+	}
+
+	if !resolvableDomainRegexp.MatchString(accountDomain) {
+		return "", fmt.Errorf("%w: account %s has domain %q, which must contain at least one dot and only "+
+			"lowercase letters, digits and hyphens. Correct the account domain before migrating",
+			ErrUnusableDomain, accountID, accountDomain)
+	}
+
+	if requested && accountDomain != singleAccountDomain {
+		return "", fmt.Errorf("%w: account %s already uses domain %q but %q was requested. Re-run without "+
+			"--single-account-mode-domain to keep %q, or correct the account domain first",
+			ErrDomainConflict, accountID, accountDomain, singleAccountDomain, accountDomain)
+	}
+
+	return accountDomain, nil
+}
+
+// EnsureSingleAccountDomain gives the remaining account the domain attributes single account mode
+// resolves against, so users can still join it after the migration.
+func EnsureSingleAccountDomain(s Server, singleAccountDomain string) error {
+	plan, err := planSingleAccountDomain(s, singleAccountDomain)
+	if err != nil {
+		return err
+	}
+	if plan.skip {
+		return nil
+	}
+
+	if isDryRun() {
+		log.Infof("[DRY RUN] would set account %s domain to %q, category to %q and mark it as the primary domain account "+
+			"(currently domain=%q primary=%v)", plan.accountID, plan.domain, types.PrivateCategory,
+			plan.currentDomain, plan.isPrimary)
+		return nil
+	}
+
+	if err := s.Store().UpdateAccountDomainAttributes(context.Background(), plan.accountID, plan.domain,
+		types.PrivateCategory, true); err != nil {
+		return fmt.Errorf("failed to update domain attributes of account %s: %w", plan.accountID, err)
+	}
+
+	log.Infof("account %s now resolves in single account mode with domain %q", plan.accountID, plan.domain)
+	return nil
+}
+
+// CheckSingleAccountDomain reports whether EnsureSingleAccountDomain would succeed, without writing.
+func CheckSingleAccountDomain(s Server, singleAccountDomain string) error {
+	_, err := planSingleAccountDomain(s, singleAccountDomain)
+	return err
+}
+
+type singleAccountDomainPlan struct {
+	accountID     string
+	domain        string
+	currentDomain string
+	isPrimary     bool
+	skip          bool
+}
+
+// planSingleAccountDomain decides what the account's domain attributes should become. It reads
+// only, so it can run both as a preflight and as the first half of the update.
+func planSingleAccountDomain(s Server, singleAccountDomain string) (singleAccountDomainPlan, error) {
+	ctx := context.Background()
+
+	// An empty value means the operator did not pick a domain, so the default is only a fallback.
+	requested := singleAccountDomain != ""
+
+	singleAccountDomain, err := NormalizeSingleAccountDomain(singleAccountDomain)
+	if err != nil {
+		return singleAccountDomainPlan{}, err
+	}
+
+	accountsCounter, err := s.Store().GetAccountsCounter(ctx)
+	if err != nil {
+		return singleAccountDomainPlan{}, fmt.Errorf("failed to count accounts: %w", err)
+	}
+	// The count is checked again here: it is read long after RequireSingleAccount, and marking an
+	// arbitrary account as the primary one for the domain would be wrong.
+	switch {
+	case accountsCounter == 0:
+		log.Info("no accounts yet, nothing to prepare for single account mode")
+		return singleAccountDomainPlan{skip: true}, nil
+	case accountsCounter > 1:
+		return singleAccountDomainPlan{}, errMultipleAccounts(accountsCounter)
+	}
+
+	accountID, err := s.Store().GetAnyAccountID(ctx)
+	if err != nil {
+		return singleAccountDomainPlan{}, fmt.Errorf("failed to get the existing account: %w", err)
+	}
+
+	isPrimary, accountDomain, err := s.Store().IsPrimaryAccount(ctx, accountID)
+	if err != nil {
+		return singleAccountDomainPlan{}, fmt.Errorf("failed to read domain attributes of account %s: %w", accountID, err)
+	}
+
+	domain, err := resolveAccountDomain(accountID, accountDomain, singleAccountDomain, requested)
+	if err != nil {
+		return singleAccountDomainPlan{}, err
+	}
+
+	return singleAccountDomainPlan{
+		accountID:     accountID,
+		domain:        domain,
+		currentDomain: accountDomain,
+		isPrimary:     isPrimary,
+	}, nil
+}

management/server/idp/migration/migration_test.go

@@ -24,6 +24,17 @@ type testStore struct {
 	checkSchemaFunc    func(checks []SchemaCheck) []SchemaError
 	updateCalls        []updateUserIDCall
 	updateInfoCalls    []updateUserInfoCall
+
+	accountsCounter int64
+	accounts        map[string]*types.Account
+	domainAttrCalls []domainAttrCall
+}
+
+type domainAttrCall struct {
+	AccountID string
+	Domain    string
+	Category  string
+	IsPrimary bool
 }
 
 type updateUserIDCall struct {
@@ -38,6 +49,35 @@ type updateUserInfoCall struct {
 	Name   string
 }
 
+func (s *testStore) GetAccountsCounter(context.Context) (int64, error) {
+	return s.accountsCounter, nil
+}
+
+func (s *testStore) GetAnyAccountID(context.Context) (string, error) {
+	for id := range s.accounts {
+		return id, nil
+	}
+	return "", fmt.Errorf("no accounts")
+}
+
+func (s *testStore) IsPrimaryAccount(_ context.Context, accountID string) (bool, string, error) {
+	account, ok := s.accounts[accountID]
+	if !ok {
+		return false, "", fmt.Errorf("account %s not found", accountID)
+	}
+	return account.IsDomainPrimaryAccount, account.Domain, nil
+}
+
+func (s *testStore) UpdateAccountDomainAttributes(_ context.Context, accountID, domain, category string, isPrimaryDomain bool) error {
+	s.domainAttrCalls = append(s.domainAttrCalls, domainAttrCall{accountID, domain, category, isPrimaryDomain})
+	if account, ok := s.accounts[accountID]; ok {
+		account.Domain = domain
+		account.DomainCategory = category
+		account.IsDomainPrimaryAccount = isPrimaryDomain
+	}
+	return nil
+}
+
 func (s *testStore) ListUsers(ctx context.Context) ([]*types.User, error) {
 	return s.listUsersFunc(ctx)
 }
@@ -826,3 +866,212 @@ func TestCheckSchema_MockStore(t *testing.T) {
 		assert.Equal(t, "email", errs[0].Column)
 	})
 }
+
+func TestRequireSingleAccount(t *testing.T) {
+	tests := []struct {
+		name      string
+		accounts  int64
+		expectErr bool
+	}{
+		{name: "fresh install", accounts: 0},
+		{name: "single account", accounts: 1},
+		{name: "multiple accounts", accounts: 3, expectErr: true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			srv := &testServer{store: &testStore{accountsCounter: tt.accounts}}
+
+			err := RequireSingleAccount(srv)
+			if !tt.expectErr {
+				require.NoError(t, err)
+				return
+			}
+
+			require.Error(t, err)
+			assert.ErrorIs(t, err, ErrMultipleAccounts)
+		})
+	}
+}
+
+func TestEnsureSingleAccountDomain(t *testing.T) {
+	tests := []struct {
+		name            string
+		account         *types.Account
+		requestedDomain string
+		expectedDomain  string
+	}{
+		{
+			name:           "account migrated from an IdP without domain claims",
+			account:        &types.Account{Id: "account-1"},
+			expectedDomain: DefaultSingleAccountDomain,
+		},
+		{
+			name:            "requested domain is applied to an account without one",
+			account:         &types.Account{Id: "account-1"},
+			requestedDomain: "corp.example.com",
+			expectedDomain:  "corp.example.com",
+		},
+		{
+			name:           "account keeps its own domain",
+			account:        &types.Account{Id: "account-1", Domain: "acme.com"},
+			expectedDomain: "acme.com",
+		},
+		{
+			name:            "requesting the domain the account already has is not a conflict",
+			account:         &types.Account{Id: "account-1", Domain: "acme.com"},
+			requestedDomain: "acme.com",
+			expectedDomain:  "acme.com",
+		},
+		{
+			name: "already resolvable account is rewritten with the same values",
+			account: &types.Account{
+				Id:                     "account-1",
+				Domain:                 "acme.com",
+				DomainCategory:         types.PrivateCategory,
+				IsDomainPrimaryAccount: true,
+			},
+			expectedDomain: "acme.com",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			store := &testStore{
+				accountsCounter: 1,
+				accounts:        map[string]*types.Account{tt.account.Id: tt.account},
+			}
+
+			require.NoError(t, EnsureSingleAccountDomain(&testServer{store: store}, tt.requestedDomain))
+
+			require.Len(t, store.domainAttrCalls, 1)
+			assert.Equal(t, domainAttrCall{
+				AccountID: tt.account.Id,
+				Domain:    tt.expectedDomain,
+				Category:  types.PrivateCategory,
+				IsPrimary: true,
+			}, store.domainAttrCalls[0])
+		})
+	}
+}
+
+func TestEnsureSingleAccountDomainDryRun(t *testing.T) {
+	t.Setenv(dryRunEnvKey, "true")
+
+	store := &testStore{
+		accountsCounter: 1,
+		accounts:        map[string]*types.Account{"account-1": {Id: "account-1"}},
+	}
+
+	require.NoError(t, EnsureSingleAccountDomain(&testServer{store: store}, ""))
+	assert.Empty(t, store.domainAttrCalls, "Dry run must not write anything")
+}
+
+func TestEnsureSingleAccountDomainRejectsUnresolvableDomains(t *testing.T) {
+	t.Run("account domain that cannot resolve is reported", func(t *testing.T) {
+		account := &types.Account{Id: "account-1", Domain: "corp"}
+		store := &testStore{
+			accountsCounter: 1,
+			accounts:        map[string]*types.Account{account.Id: account},
+		}
+
+		err := EnsureSingleAccountDomain(&testServer{store: store}, "")
+		require.Error(t, err)
+		assert.ErrorIs(t, err, ErrUnusableDomain)
+		assert.Empty(t, store.domainAttrCalls, "A broken account domain must not be replaced silently")
+	})
+
+	t.Run("requested domain conflicting with the account domain is reported", func(t *testing.T) {
+		account := &types.Account{Id: "account-1", Domain: "acme.com"}
+		store := &testStore{
+			accountsCounter: 1,
+			accounts:        map[string]*types.Account{account.Id: account},
+		}
+
+		err := EnsureSingleAccountDomain(&testServer{store: store}, "corp.example.com")
+		require.Error(t, err)
+		assert.ErrorIs(t, err, ErrDomainConflict)
+		assert.Empty(t, store.domainAttrCalls, "A conflict must not overwrite the account domain")
+	})
+
+	t.Run("configured domain that cannot resolve is rejected", func(t *testing.T) {
+		store := &testStore{
+			accountsCounter: 1,
+			accounts:        map[string]*types.Account{"account-1": {Id: "account-1"}},
+		}
+
+		err := EnsureSingleAccountDomain(&testServer{store: store}, "corp")
+		require.Error(t, err)
+		assert.ErrorIs(t, err, ErrUnusableDomain)
+		assert.Empty(t, store.domainAttrCalls)
+	})
+
+	t.Run("account appearing after the preflight is rejected", func(t *testing.T) {
+		store := &testStore{
+			accountsCounter: 2,
+			accounts: map[string]*types.Account{
+				"account-1": {Id: "account-1"},
+				"account-2": {Id: "account-2"},
+			},
+		}
+
+		err := EnsureSingleAccountDomain(&testServer{store: store}, "")
+		require.Error(t, err)
+		assert.ErrorIs(t, err, ErrMultipleAccounts)
+		assert.Empty(t, store.domainAttrCalls, "No account may be marked primary when several exist")
+	})
+
+	t.Run("fresh install with no accounts is a no-op", func(t *testing.T) {
+		store := &testStore{accountsCounter: 0, accounts: map[string]*types.Account{}}
+
+		require.NoError(t, EnsureSingleAccountDomain(&testServer{store: store}, ""))
+		assert.Empty(t, store.domainAttrCalls)
+	})
+}
+
+func TestCheckSingleAccountDomain(t *testing.T) {
+	tests := []struct {
+		name      string
+		account   *types.Account
+		requested string
+		expectErr error
+	}{
+		{
+			name:    "usable account domain passes",
+			account: &types.Account{Id: "account-1", Domain: "acme.com"},
+		},
+		{
+			name:    "empty account domain passes",
+			account: &types.Account{Id: "account-1"},
+		},
+		{
+			name:      "unresolvable account domain fails",
+			account:   &types.Account{Id: "account-1", Domain: "corp"},
+			expectErr: ErrUnusableDomain,
+		},
+		{
+			name:      "conflicting request fails",
+			account:   &types.Account{Id: "account-1", Domain: "acme.com"},
+			requested: "corp.example.com",
+			expectErr: ErrDomainConflict,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			store := &testStore{
+				accountsCounter: 1,
+				accounts:        map[string]*types.Account{tt.account.Id: tt.account},
+			}
+
+			err := CheckSingleAccountDomain(&testServer{store: store}, tt.requested)
+			if tt.expectErr == nil {
+				require.NoError(t, err)
+			} else {
+				require.ErrorIs(t, err, tt.expectErr)
+			}
+
+			assert.Empty(t, store.domainAttrCalls, "The preflight must not write anything")
+		})
+	}
+}

management/server/idp/migration/store.go

@@ -60,6 +60,20 @@ type Store interface {
 	// CheckSchema verifies that all tables and columns required by the migration
 	// exist in the database. Returns a list of problems; an empty slice means OK.
 	CheckSchema(checks []SchemaCheck) []SchemaError
+
+	// GetAccountsCounter returns the total number of accounts in the store.
+	GetAccountsCounter(ctx context.Context) (int64, error)
+
+	// GetAnyAccountID returns the ID of one of the existing accounts.
+	GetAnyAccountID(ctx context.Context) (string, error)
+
+	// IsPrimaryAccount returns whether the account is the primary account for its domain,
+	// along with that domain.
+	IsPrimaryAccount(ctx context.Context, accountID string) (bool, string, error)
+
+	// UpdateAccountDomainAttributes sets the domain, domain category and primary
+	// domain flag of an account.
+	UpdateAccountDomainAttributes(ctx context.Context, accountID string, domain string, category string, isPrimaryDomain bool) error
 }
 
 // RequiredEventSchema lists all tables and columns that the migration tool needs

management/server/idp/util.go

@@ -2,11 +2,12 @@ package idp
 
 import (
 	"encoding/json"
-	"math/rand"
 	"net/url"
 	"os"
 	"strings"
 	"time"
+
+	"github.com/netbirdio/netbird/management/server/util"
 )
 
 var (
@@ -33,31 +34,32 @@ func GeneratePassword(passwordLength, minSpecialChar, minNum, minUpperCase int)
 
 	//Set special character
 	for i := 0; i < minSpecialChar; i++ {
-		random := rand.Intn(len(specialCharSet))
+		random := util.RandIntn(len(specialCharSet))
 		password.WriteString(string(specialCharSet[random]))
 	}
 
 	//Set numeric
 	for i := 0; i < minNum; i++ {
-		random := rand.Intn(len(numberSet))
+		random := util.RandIntn(len(numberSet))
 		password.WriteString(string(numberSet[random]))
 	}
 
 	//Set uppercase
 	for i := 0; i < minUpperCase; i++ {
-		random := rand.Intn(len(upperCharSet))
+		random := util.RandIntn(len(upperCharSet))
 		password.WriteString(string(upperCharSet[random]))
 	}
 
 	remainingLength := passwordLength - minSpecialChar - minNum - minUpperCase
 	for i := 0; i < remainingLength; i++ {
-		random := rand.Intn(len(allCharSet))
+		random := util.RandIntn(len(allCharSet))
 		password.WriteString(string(allCharSet[random]))
 	}
 	inRune := []rune(password.String())
-	rand.Shuffle(len(inRune), func(i, j int) {
+	for i := len(inRune) - 1; i > 0; i-- {
+		j := util.RandIntn(i + 1)
 		inRune[i], inRune[j] = inRune[j], inRune[i]
-	})
+	}
 	return string(inRune)
 }
 

management/server/instance/manager.go

@@ -12,6 +12,7 @@ import (
 	"sync"
 	"time"
 
+	"github.com/dexidp/dex/storage"
 	goversion "github.com/hashicorp/go-version"
 	log "github.com/sirupsen/logrus"
 
@@ -60,6 +61,13 @@ type Manager interface {
 	// This should only be called when IsSetupRequired returns true.
 	CreateOwnerUser(ctx context.Context, email, password, name string) (*idp.UserData, error)
 
+	// RollbackSetup reverses a successful CreateOwnerUser by deleting the user
+	// from the embedded IDP and reloading setupRequired from persistent state, so
+	// /api/setup can be retried only when no accounts or local users remain. Used
+	// when post-user steps (account or PAT creation) fail and the caller wants a
+	// clean slate.
+	RollbackSetup(ctx context.Context, userID string) error
+
 	// GetVersionInfo returns version information for NetBird components.
 	GetVersionInfo(ctx context.Context) (*VersionInfo, error)
 }
@@ -70,6 +78,7 @@ type instanceStore interface {
 
 type embeddedIdP interface {
 	CreateUserWithPassword(ctx context.Context, email, password, name string) (*idp.UserData, error)
+	DeleteUser(ctx context.Context, userID string) error
 	GetAllAccounts(ctx context.Context) (map[string][]*idp.UserData, error)
 }
 
@@ -187,6 +196,51 @@ func (m *DefaultManager) CreateOwnerUser(ctx context.Context, email, password, n
 	return userData, nil
 }
 
+// RollbackSetup undoes a successful CreateOwnerUser: deletes the user from the
+// embedded IDP and reloads setupRequired from persistent state.
+func (m *DefaultManager) RollbackSetup(ctx context.Context, userID string) error {
+	if m.embeddedIdpManager == nil {
+		return errors.New("embedded IDP is not enabled")
+	}
+
+	var deleteErr error
+	if err := m.embeddedIdpManager.DeleteUser(ctx, userID); err != nil {
+		if isNotFoundError(err) {
+			log.WithContext(ctx).Debugf("setup rollback user %s already deleted", userID)
+		} else {
+			deleteErr = fmt.Errorf("failed to delete user from embedded IdP: %w", err)
+		}
+	}
+
+	if err := m.loadSetupRequired(ctx); err != nil {
+		reloadErr := fmt.Errorf("failed to reload setup state after rollback: %w", err)
+		if deleteErr != nil {
+			return errors.Join(deleteErr, reloadErr)
+		}
+		return reloadErr
+	}
+
+	if deleteErr != nil {
+		return deleteErr
+	}
+
+	log.WithContext(ctx).Infof("rolled back setup for user %s", userID)
+	return nil
+}
+
+func isNotFoundError(err error) bool {
+	if err == nil {
+		return false
+	}
+	if errors.Is(err, storage.ErrNotFound) {
+		return true
+	}
+	if s, ok := status.FromError(err); ok {
+		return s.Type() == status.NotFound
+	}
+	return false
+}
+
 func (m *DefaultManager) checkSetupRequiredFromDB(ctx context.Context) error {
 	numAccounts, err := m.store.GetAccountsCounter(ctx)
 	if err != nil {

management/server/instance/manager_test.go

@@ -10,15 +10,18 @@ import (
 	"testing"
 	"time"
 
+	"github.com/dexidp/dex/storage"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
 	"github.com/netbirdio/netbird/management/server/idp"
+	"github.com/netbirdio/netbird/shared/management/status"
 )
 
 type mockIdP struct {
 	mu                sync.Mutex
 	createUserFunc    func(ctx context.Context, email, password, name string) (*idp.UserData, error)
+	deleteUserFunc    func(ctx context.Context, userID string) error
 	users             map[string][]*idp.UserData
 	getAllAccountsErr error
 }
@@ -30,6 +33,13 @@ func (m *mockIdP) CreateUserWithPassword(ctx context.Context, email, password, n
 	return &idp.UserData{ID: "test-user-id", Email: email, Name: name}, nil
 }
 
+func (m *mockIdP) DeleteUser(ctx context.Context, userID string) error {
+	if m.deleteUserFunc != nil {
+		return m.deleteUserFunc(ctx, userID)
+	}
+	return nil
+}
+
 func (m *mockIdP) GetAllAccounts(_ context.Context) (map[string][]*idp.UserData, error) {
 	m.mu.Lock()
 	defer m.mu.Unlock()
@@ -223,6 +233,77 @@ func TestIsSetupRequired_ReturnsFlag(t *testing.T) {
 	assert.False(t, required)
 }
 
+func TestRollbackSetup_UserAlreadyDeletedIsSuccess(t *testing.T) {
+	tests := []struct {
+		name string
+		err  error
+	}{
+		{
+			name: "management status not found",
+			err:  status.NewUserNotFoundError("owner-id"),
+		},
+		{
+			name: "dex storage not found",
+			err:  fmt.Errorf("failed to get user for deletion: %w", storage.ErrNotFound),
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			idpMock := &mockIdP{
+				deleteUserFunc: func(_ context.Context, userID string) error {
+					assert.Equal(t, "owner-id", userID)
+					return tt.err
+				},
+			}
+			mgr := newTestManager(idpMock, &mockStore{})
+			mgr.setupRequired = false
+
+			err := mgr.RollbackSetup(context.Background(), "owner-id")
+			require.NoError(t, err)
+
+			required, err := mgr.IsSetupRequired(context.Background())
+			require.NoError(t, err)
+			assert.True(t, required, "setup should be required when no accounts or local users remain")
+		})
+	}
+}
+
+func TestRollbackSetup_RecomputesSetupStateWhenAccountStillExists(t *testing.T) {
+	idpMock := &mockIdP{
+		deleteUserFunc: func(_ context.Context, _ string) error {
+			return status.NewUserNotFoundError("owner-id")
+		},
+	}
+	mgr := newTestManager(idpMock, &mockStore{accountsCount: 1})
+	mgr.setupRequired = true
+
+	err := mgr.RollbackSetup(context.Background(), "owner-id")
+	require.NoError(t, err)
+
+	required, err := mgr.IsSetupRequired(context.Background())
+	require.NoError(t, err)
+	assert.False(t, required, "setup should not be required while an account still exists")
+}
+
+func TestRollbackSetup_ReturnsDeleteErrorButReloadsSetupState(t *testing.T) {
+	idpMock := &mockIdP{
+		deleteUserFunc: func(_ context.Context, _ string) error {
+			return errors.New("idp unavailable")
+		},
+	}
+	mgr := newTestManager(idpMock, &mockStore{})
+	mgr.setupRequired = false
+
+	err := mgr.RollbackSetup(context.Background(), "owner-id")
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "idp unavailable")
+
+	required, err := mgr.IsSetupRequired(context.Background())
+	require.NoError(t, err)
+	assert.True(t, required, "setup state should be reloaded even when user deletion fails")
+}
+
 func TestDefaultManager_ValidateSetupRequest(t *testing.T) {
 	manager := &DefaultManager{setupRequired: true}
 

management/server/instance/setup_service.go

@@ -0,0 +1,216 @@
+package instance
+
+import (
+	"context"
+	"fmt"
+	"os"
+
+	log "github.com/sirupsen/logrus"
+
+	"github.com/netbirdio/netbird/management/server/account"
+	"github.com/netbirdio/netbird/management/server/idp"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/shared/auth"
+	"github.com/netbirdio/netbird/shared/management/status"
+)
+
+const (
+	setupPATTokenName = "setup-token"
+
+	// SetupPATEnabledEnvKey enables setup-time Personal Access Token creation.
+	SetupPATEnabledEnvKey = "NB_SETUP_PAT_ENABLED"
+
+	setupPATDefaultExpireDays = 1
+)
+
+// SetupOptions controls optional work performed during initial instance setup.
+type SetupOptions struct {
+	// CreatePAT requests creation of a setup Personal Access Token. It is honored
+	// only when SetupPATEnabledEnvKey is set to "true".
+	CreatePAT bool
+	// PATExpireInDays defaults to 1 day when CreatePAT is requested and setup PAT
+	// creation is enabled.
+	PATExpireInDays *int
+}
+
+// SetupResult contains resources created during initial instance setup.
+type SetupResult struct {
+	User          *idp.UserData
+	PATPlainToken string
+}
+
+// SetupService orchestrates the initial setup use case across the instance and
+// account bounded contexts and owns the compensation logic when a later step
+// fails.
+type SetupService struct {
+	instanceManager Manager
+	accountManager  account.Manager
+	setupPATEnabled bool
+}
+
+// NewSetupService creates a setup use-case service.
+func NewSetupService(instanceManager Manager, accountManager account.Manager) *SetupService {
+	return &SetupService{
+		instanceManager: instanceManager,
+		accountManager:  accountManager,
+		setupPATEnabled: os.Getenv(SetupPATEnabledEnvKey) == "true",
+	}
+}
+
+func normalizeSetupOptions(opts SetupOptions, setupPATEnabled bool) (SetupOptions, error) {
+	if !opts.CreatePAT {
+		return opts, nil
+	}
+
+	if !setupPATEnabled {
+		opts.CreatePAT = false
+		opts.PATExpireInDays = nil
+		return opts, nil
+	}
+
+	if opts.PATExpireInDays == nil {
+		defaultExpireInDays := setupPATDefaultExpireDays
+		opts.PATExpireInDays = &defaultExpireInDays
+	}
+
+	if *opts.PATExpireInDays < account.PATMinExpireDays || *opts.PATExpireInDays > account.PATMaxExpireDays {
+		return opts, status.Errorf(status.InvalidArgument, "pat_expire_in must be between %d and %d", account.PATMinExpireDays, account.PATMaxExpireDays)
+	}
+
+	return opts, nil
+}
+
+// SetupOwner creates the initial owner user and, when requested and enabled by
+// SetupPATEnabledEnvKey, provisions the account and a setup Personal Access
+// Token. If account or PAT provisioning fails, created resources are rolled
+// back so setup can be retried. If account rollback fails, user rollback is
+// skipped to avoid leaving an account without its owner user.
+func (m *SetupService) SetupOwner(ctx context.Context, email, password, name string, opts SetupOptions) (*SetupResult, error) {
+	opts, err := normalizeSetupOptions(opts, m.setupPATEnabled)
+	if err != nil {
+		return nil, err
+	}
+
+	if opts.CreatePAT && m.accountManager == nil {
+		return nil, fmt.Errorf("account manager is required to create setup PAT")
+	}
+
+	userData, err := m.instanceManager.CreateOwnerUser(ctx, email, password, name)
+	if err != nil {
+		return nil, err
+	}
+
+	result := &SetupResult{User: userData}
+	if !opts.CreatePAT {
+		return result, nil
+	}
+
+	userAuth := auth.UserAuth{
+		UserId: userData.ID,
+		Email:  userData.Email,
+		Name:   userData.Name,
+	}
+
+	accountID, err := m.accountManager.GetAccountIDByUserID(ctx, userAuth)
+	if err != nil {
+		err = fmt.Errorf("create account for setup user: %w", err)
+		if rollbackErr := m.rollbackSetup(ctx, userData.ID, "account provisioning failed", err, ""); rollbackErr != nil {
+			return nil, fmt.Errorf("%w; failed to roll back setup resources: %v", err, rollbackErr)
+		}
+		return nil, err
+	}
+
+	pat, err := m.accountManager.CreatePAT(ctx, accountID, userData.ID, userData.ID, setupPATTokenName, *opts.PATExpireInDays)
+	if err != nil {
+		err = fmt.Errorf("create setup PAT: %w", err)
+		if rollbackErr := m.rollbackSetup(ctx, userData.ID, "setup PAT provisioning failed", err, accountID); rollbackErr != nil {
+			return nil, fmt.Errorf("%w; failed to roll back setup resources: %v", err, rollbackErr)
+		}
+		return nil, err
+	}
+
+	result.PATPlainToken = pat.PlainToken
+	return result, nil
+}
+
+func (m *SetupService) rollbackSetup(ctx context.Context, userID, reason string, origErr error, accountID string) error {
+	if accountID == "" {
+		resolvedAccountID, err := m.lookupSetupAccountIDForRollback(ctx, userID)
+		if err != nil {
+			rollbackErr := fmt.Errorf("resolve setup account for rollback: %w", err)
+			log.WithContext(ctx).Errorf("failed to resolve setup account for user %s after %s: original error: %v, rollback error: %v", userID, reason, origErr, rollbackErr)
+			return rollbackErr
+		}
+		accountID = resolvedAccountID
+	}
+
+	if accountID != "" {
+		if err := m.rollbackSetupAccount(ctx, accountID); err != nil {
+			rollbackErr := fmt.Errorf("roll back setup account %s: %w", accountID, err)
+			log.WithContext(ctx).Errorf("failed to roll back setup account %s for user %s after %s: original error: %v, rollback error: %v", accountID, userID, reason, origErr, rollbackErr)
+			return rollbackErr
+		}
+		log.WithContext(ctx).Warnf("rolled back setup account %s for user %s after %s: %v", accountID, userID, reason, origErr)
+	}
+
+	if err := m.instanceManager.RollbackSetup(ctx, userID); err != nil {
+		rollbackErr := fmt.Errorf("roll back setup user %s: %w", userID, err)
+		log.WithContext(ctx).Errorf("failed to roll back setup user %s after %s: original error: %v, rollback error: %v", userID, reason, origErr, rollbackErr)
+		return rollbackErr
+	}
+	log.WithContext(ctx).Warnf("rolled back setup user %s after %s: %v", userID, reason, origErr)
+	return nil
+}
+
+func (m *SetupService) lookupSetupAccountIDForRollback(ctx context.Context, userID string) (string, error) {
+	if m.accountManager == nil {
+		return "", fmt.Errorf("account manager is required to resolve setup account")
+	}
+
+	accountStore := m.accountManager.GetStore()
+	if accountStore == nil {
+		return "", fmt.Errorf("account store is unavailable")
+	}
+
+	accountID, err := accountStore.GetAccountIDByUserID(ctx, store.LockingStrengthNone, userID)
+	if err != nil {
+		if isNotFoundError(err) {
+			return "", nil
+		}
+		return "", fmt.Errorf("get setup account ID for rollback: %w", err)
+	}
+
+	return accountID, nil
+}
+
+// rollbackSetupAccount removes only the setup-created account data from the
+// store. It intentionally avoids accountManager.DeleteAccount because the normal
+// account deletion path also deletes users from the IdP; embedded IdP cleanup is
+// owned by instanceManager.RollbackSetup.
+func (m *SetupService) rollbackSetupAccount(ctx context.Context, accountID string) error {
+	if m.accountManager == nil {
+		return fmt.Errorf("account manager is required to roll back setup account")
+	}
+
+	accountStore := m.accountManager.GetStore()
+	if accountStore == nil {
+		return fmt.Errorf("account store is unavailable")
+	}
+
+	account, err := accountStore.GetAccount(ctx, accountID)
+	if err != nil {
+		if isNotFoundError(err) {
+			return nil
+		}
+		return fmt.Errorf("get setup account for rollback: %w", err)
+	}
+
+	if err := accountStore.DeleteAccount(ctx, account); err != nil {
+		if isNotFoundError(err) {
+			return nil
+		}
+		return fmt.Errorf("delete setup account for rollback: %w", err)
+	}
+
+	return nil
+}

management/server/instance/setup_service_test.go

@@ -0,0 +1,318 @@
+package instance
+
+import (
+	"context"
+	"errors"
+	"testing"
+
+	"go.uber.org/mock/gomock"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/server/idp"
+	"github.com/netbirdio/netbird/management/server/mock_server"
+	nbstore "github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/auth"
+	"github.com/netbirdio/netbird/shared/management/status"
+)
+
+type setupInstanceManagerMock struct {
+	createOwnerUserFn func(ctx context.Context, email, password, name string) (*idp.UserData, error)
+	rollbackSetupFn   func(ctx context.Context, userID string) error
+}
+
+func (m *setupInstanceManagerMock) IsSetupRequired(context.Context) (bool, error) {
+	return true, nil
+}
+
+func (m *setupInstanceManagerMock) CreateOwnerUser(ctx context.Context, email, password, name string) (*idp.UserData, error) {
+	if m.createOwnerUserFn != nil {
+		return m.createOwnerUserFn(ctx, email, password, name)
+	}
+	return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+}
+
+func (m *setupInstanceManagerMock) RollbackSetup(ctx context.Context, userID string) error {
+	if m.rollbackSetupFn != nil {
+		return m.rollbackSetupFn(ctx, userID)
+	}
+	return nil
+}
+
+func (m *setupInstanceManagerMock) GetVersionInfo(context.Context) (*VersionInfo, error) {
+	return &VersionInfo{}, nil
+}
+
+var _ Manager = (*setupInstanceManagerMock)(nil)
+
+func intPtr(v int) *int {
+	return &v
+}
+
+func TestSetupOwner_PATFeatureDisabled_IgnoresCreatePAT(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "false")
+
+	createCalls := 0
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			createOwnerUserFn: func(_ context.Context, email, _, name string) (*idp.UserData, error) {
+				createCalls++
+				return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+			},
+		},
+		&mock_server.MockAccountManager{},
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT: true,
+	})
+
+	require.NoError(t, err)
+	require.NotNil(t, result)
+	assert.Equal(t, "owner-id", result.User.ID)
+	assert.Empty(t, result.PATPlainToken)
+	assert.Equal(t, 1, createCalls)
+}
+
+func TestSetupOwner_PATFeatureEnabled_MissingExpireDefaultsToOneDay(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "true")
+
+	createCalled := false
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			createOwnerUserFn: func(_ context.Context, email, _, name string) (*idp.UserData, error) {
+				createCalled = true
+				return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+			},
+		},
+		&mock_server.MockAccountManager{
+			GetAccountIDByUserIdFunc: func(_ context.Context, userAuth auth.UserAuth) (string, error) {
+				assert.Equal(t, "owner-id", userAuth.UserId)
+				return "acc-1", nil
+			},
+			CreatePATFunc: func(_ context.Context, accountID, initiatorUserID, targetUserID, tokenName string, expiresIn int) (*types.PersonalAccessTokenGenerated, error) {
+				assert.Equal(t, "acc-1", accountID)
+				assert.Equal(t, "owner-id", initiatorUserID)
+				assert.Equal(t, "owner-id", targetUserID)
+				assert.Equal(t, setupPATTokenName, tokenName)
+				assert.Equal(t, setupPATDefaultExpireDays, expiresIn)
+				return &types.PersonalAccessTokenGenerated{PlainToken: "nbp_plain"}, nil
+			},
+		},
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT: true,
+	})
+
+	require.NoError(t, err)
+	require.NotNil(t, result)
+	assert.True(t, createCalled)
+	assert.Equal(t, "nbp_plain", result.PATPlainToken)
+}
+
+func TestSetupOwner_PATFeatureEnabled_MissingAccountManagerFailsBeforeCreateUser(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "true")
+
+	createCalled := false
+	rollbackCalled := false
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			createOwnerUserFn: func(_ context.Context, email, _, name string) (*idp.UserData, error) {
+				createCalled = true
+				return &idp.UserData{ID: "owner-id", Email: email, Name: name}, nil
+			},
+			rollbackSetupFn: func(_ context.Context, _ string) error {
+				rollbackCalled = true
+				return nil
+			},
+		},
+		nil,
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT: true,
+	})
+
+	require.Error(t, err)
+	assert.Nil(t, result)
+	assert.Contains(t, err.Error(), "account manager is required")
+	assert.False(t, createCalled)
+	assert.False(t, rollbackCalled)
+}
+
+func TestSetupOwner_AccountProvisioningFails_RollsBackSideEffectAccountAndUser(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "true")
+
+	ctrl := gomock.NewController(t)
+	accountStore := nbstore.NewMockStore(ctrl)
+	account := &types.Account{Id: "acc-1"}
+	accountStore.EXPECT().GetAccountIDByUserID(gomock.Any(), nbstore.LockingStrengthNone, "owner-id").Return("acc-1", nil)
+	accountStore.EXPECT().GetAccount(gomock.Any(), "acc-1").Return(account, nil)
+	accountStore.EXPECT().DeleteAccount(gomock.Any(), account).Return(nil)
+
+	rolledBackFor := ""
+	rollbackCalls := 0
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			rollbackSetupFn: func(_ context.Context, userID string) error {
+				rollbackCalls++
+				rolledBackFor = userID
+				return nil
+			},
+		},
+		&mock_server.MockAccountManager{
+			GetAccountIDByUserIdFunc: func(_ context.Context, userAuth auth.UserAuth) (string, error) {
+				assert.Equal(t, "owner-id", userAuth.UserId)
+				return "", errors.New("metadata update failed")
+			},
+			GetStoreFunc: func() nbstore.Store {
+				return accountStore
+			},
+		},
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT:       true,
+		PATExpireInDays: intPtr(30),
+	})
+
+	require.Error(t, err)
+	assert.Nil(t, result)
+	assert.Contains(t, err.Error(), "create account for setup user")
+	assert.Equal(t, "owner-id", rolledBackFor)
+	assert.Equal(t, 1, rollbackCalls)
+}
+
+func TestSetupOwner_CreatePATFails_RollsBackSetupAccountAndUser(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "true")
+
+	ctrl := gomock.NewController(t)
+	accountStore := nbstore.NewMockStore(ctrl)
+	account := &types.Account{Id: "acc-1"}
+	accountStore.EXPECT().GetAccount(gomock.Any(), "acc-1").Return(account, nil)
+	accountStore.EXPECT().DeleteAccount(gomock.Any(), account).Return(nil)
+
+	rollbackCalls := 0
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			rollbackSetupFn: func(_ context.Context, userID string) error {
+				rollbackCalls++
+				assert.Equal(t, "owner-id", userID)
+				return nil
+			},
+		},
+		&mock_server.MockAccountManager{
+			GetAccountIDByUserIdFunc: func(_ context.Context, userAuth auth.UserAuth) (string, error) {
+				assert.Equal(t, "owner-id", userAuth.UserId)
+				return "acc-1", nil
+			},
+			CreatePATFunc: func(_ context.Context, accountID, initiatorUserID, targetUserID, tokenName string, expiresIn int) (*types.PersonalAccessTokenGenerated, error) {
+				assert.Equal(t, "acc-1", accountID)
+				assert.Equal(t, "owner-id", initiatorUserID)
+				assert.Equal(t, "owner-id", targetUserID)
+				assert.Equal(t, setupPATTokenName, tokenName)
+				assert.Equal(t, 30, expiresIn)
+				return nil, status.Errorf(status.Internal, "token store unavailable")
+			},
+			GetStoreFunc: func() nbstore.Store {
+				return accountStore
+			},
+		},
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT:       true,
+		PATExpireInDays: intPtr(30),
+	})
+
+	require.Error(t, err)
+	assert.Nil(t, result)
+	assert.Contains(t, err.Error(), "create setup PAT")
+	assert.Equal(t, 1, rollbackCalls)
+}
+
+func TestSetupOwner_CreatePATFails_AccountAlreadyGoneStillRollsBackUser(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "true")
+
+	ctrl := gomock.NewController(t)
+	accountStore := nbstore.NewMockStore(ctrl)
+	accountStore.EXPECT().GetAccount(gomock.Any(), "acc-1").Return(nil, status.NewAccountNotFoundError("acc-1"))
+
+	rolledBackFor := ""
+	rollbackCalls := 0
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			rollbackSetupFn: func(_ context.Context, userID string) error {
+				rollbackCalls++
+				rolledBackFor = userID
+				return nil
+			},
+		},
+		&mock_server.MockAccountManager{
+			GetAccountIDByUserIdFunc: func(_ context.Context, _ auth.UserAuth) (string, error) {
+				return "acc-1", nil
+			},
+			CreatePATFunc: func(_ context.Context, _, _, _, _ string, _ int) (*types.PersonalAccessTokenGenerated, error) {
+				return nil, errors.New("token failure")
+			},
+			GetStoreFunc: func() nbstore.Store {
+				return accountStore
+			},
+		},
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT:       true,
+		PATExpireInDays: intPtr(30),
+	})
+
+	require.Error(t, err)
+	assert.Nil(t, result)
+	assert.Contains(t, err.Error(), "create setup PAT")
+	assert.Equal(t, "owner-id", rolledBackFor)
+	assert.Equal(t, 1, rollbackCalls)
+}
+
+func TestSetupOwner_CreatePATFails_AccountRollbackFailureStopsBeforeUserRollback(t *testing.T) {
+	t.Setenv(SetupPATEnabledEnvKey, "true")
+
+	ctrl := gomock.NewController(t)
+	accountStore := nbstore.NewMockStore(ctrl)
+	account := &types.Account{Id: "acc-1"}
+	accountStore.EXPECT().GetAccount(gomock.Any(), "acc-1").Return(account, nil)
+	accountStore.EXPECT().DeleteAccount(gomock.Any(), account).Return(errors.New("delete failed"))
+
+	rollbackCalls := 0
+	setupManager := NewSetupService(
+		&setupInstanceManagerMock{
+			rollbackSetupFn: func(_ context.Context, userID string) error {
+				rollbackCalls++
+				return nil
+			},
+		},
+		&mock_server.MockAccountManager{
+			GetAccountIDByUserIdFunc: func(_ context.Context, _ auth.UserAuth) (string, error) {
+				return "acc-1", nil
+			},
+			CreatePATFunc: func(_ context.Context, _, _, _, _ string, _ int) (*types.PersonalAccessTokenGenerated, error) {
+				return nil, errors.New("token failure")
+			},
+			GetStoreFunc: func() nbstore.Store {
+				return accountStore
+			},
+		},
+	)
+
+	result, err := setupManager.SetupOwner(context.Background(), "admin@example.com", "securepassword123", "Admin", SetupOptions{
+		CreatePAT:       true,
+		PATExpireInDays: intPtr(30),
+	})
+
+	require.Error(t, err)
+	assert.Nil(t, result)
+	assert.Contains(t, err.Error(), "create setup PAT")
+	assert.Contains(t, err.Error(), "failed to roll back setup resources")
+	assert.Equal(t, 0, rollbackCalls)
+}

management/server/integrated_validator.go

@@ -11,6 +11,7 @@ import (
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/management/networkmap/nmdata"
 )
 
 // UpdateIntegratedValidator updates the integrated validator groups for a specified account.
@@ -109,7 +110,7 @@ func (am *DefaultAccountManager) GetValidatedPeers(ctx context.Context, accountI
 		return nil, nil, err
 	}
 
-	validPeers, err := am.integratedPeerValidator.GetValidatedPeers(ctx, accountID, groups, peers, settings.Extra)
+	validPeers, err := am.integratedPeerValidator.GetValidatedPeers(ctx, accountID, types.TwinGroups(groups), types.TwinPeers(peers), settings.Extra)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -138,7 +139,7 @@ func (a MockIntegratedValidator) ValidatePeer(_ context.Context, update *nbpeer.
 	return update, false, nil
 }
 
-func (a MockIntegratedValidator) GetValidatedPeers(_ context.Context, accountID string, groups []*types.Group, peers []*nbpeer.Peer, extraSettings *types.ExtraSettings) (map[string]struct{}, error) {
+func (a MockIntegratedValidator) GetValidatedPeers(_ context.Context, accountID string, groups []*nmdata.Group, peers []*nmdata.Peer, extraSettings *types.ExtraSettings) (map[string]struct{}, error) {
 	validatedPeers := make(map[string]struct{})
 	for _, peer := range peers {
 		validatedPeers[peer.ID] = struct{}{}

management/server/integration_reference/integration_reference.go

@@ -1,23 +0,0 @@
-package integration_reference
-
-import (
-	"fmt"
-	"strings"
-)
-
-// IntegrationReference holds the reference to a particular integration
-type IntegrationReference struct {
-	ID              int
-	IntegrationType string
-}
-
-func (ir IntegrationReference) String() string {
-	return fmt.Sprintf("%s:%d", ir.IntegrationType, ir.ID)
-}
-
-func (ir IntegrationReference) CacheKey(path ...string) string {
-	if len(path) == 0 {
-		return ir.String()
-	}
-	return fmt.Sprintf("%s:%s", ir.String(), strings.Join(path, ":"))
-}

management/server/integrations/integrated_validator/integrated_validator_mock.go

@@ -0,0 +1,187 @@
+// Code generated by MockGen. DO NOT EDIT.
+// Source: ./interface.go
+//
+// Generated by this command:
+//
+//	mockgen -package integrated_validator -destination=integrated_validator_mock.go -source=./interface.go -build_flags=-mod=mod
+//
+
+// Package integrated_validator is a generated GoMock package.
+package integrated_validator
+
+import (
+	context "context"
+	reflect "reflect"
+
+	peer "github.com/netbirdio/netbird/management/server/peer"
+	types "github.com/netbirdio/netbird/management/server/types"
+	nmdata "github.com/netbirdio/netbird/shared/management/networkmap/nmdata"
+	proto "github.com/netbirdio/netbird/shared/management/proto"
+	gomock "go.uber.org/mock/gomock"
+)
+
+// MockIntegratedValidator is a mock of IntegratedValidator interface.
+type MockIntegratedValidator struct {
+	ctrl     *gomock.Controller
+	recorder *MockIntegratedValidatorMockRecorder
+	isgomock struct{}
+}
+
+// MockIntegratedValidatorMockRecorder is the mock recorder for MockIntegratedValidator.
+type MockIntegratedValidatorMockRecorder struct {
+	mock *MockIntegratedValidator
+}
+
+// NewMockIntegratedValidator creates a new mock instance.
+func NewMockIntegratedValidator(ctrl *gomock.Controller) *MockIntegratedValidator {
+	mock := &MockIntegratedValidator{ctrl: ctrl}
+	mock.recorder = &MockIntegratedValidatorMockRecorder{mock}
+	return mock
+}
+
+// EXPECT returns an object that allows the caller to indicate expected use.
+func (m *MockIntegratedValidator) EXPECT() *MockIntegratedValidatorMockRecorder {
+	return m.recorder
+}
+
+// GetInvalidPeers mocks base method.
+func (m *MockIntegratedValidator) GetInvalidPeers(ctx context.Context, accountID string, extraSettings *types.ExtraSettings) (map[string]string, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetInvalidPeers", ctx, accountID, extraSettings)
+	ret0, _ := ret[0].(map[string]string)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetInvalidPeers indicates an expected call of GetInvalidPeers.
+func (mr *MockIntegratedValidatorMockRecorder) GetInvalidPeers(ctx, accountID, extraSettings any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetInvalidPeers", reflect.TypeOf((*MockIntegratedValidator)(nil).GetInvalidPeers), ctx, accountID, extraSettings)
+}
+
+// GetValidatedPeers mocks base method.
+func (m *MockIntegratedValidator) GetValidatedPeers(ctx context.Context, accountID string, groups []*nmdata.Group, peers []*nmdata.Peer, extraSettings *types.ExtraSettings) (map[string]struct{}, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetValidatedPeers", ctx, accountID, groups, peers, extraSettings)
+	ret0, _ := ret[0].(map[string]struct{})
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetValidatedPeers indicates an expected call of GetValidatedPeers.
+func (mr *MockIntegratedValidatorMockRecorder) GetValidatedPeers(ctx, accountID, groups, peers, extraSettings any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetValidatedPeers", reflect.TypeOf((*MockIntegratedValidator)(nil).GetValidatedPeers), ctx, accountID, groups, peers, extraSettings)
+}
+
+// IsNotValidPeer mocks base method.
+func (m *MockIntegratedValidator) IsNotValidPeer(ctx context.Context, accountID string, arg2 *peer.Peer, peersGroup []string, extraSettings *types.ExtraSettings) (bool, bool, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "IsNotValidPeer", ctx, accountID, arg2, peersGroup, extraSettings)
+	ret0, _ := ret[0].(bool)
+	ret1, _ := ret[1].(bool)
+	ret2, _ := ret[2].(error)
+	return ret0, ret1, ret2
+}
+
+// IsNotValidPeer indicates an expected call of IsNotValidPeer.
+func (mr *MockIntegratedValidatorMockRecorder) IsNotValidPeer(ctx, accountID, arg2, peersGroup, extraSettings any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "IsNotValidPeer", reflect.TypeOf((*MockIntegratedValidator)(nil).IsNotValidPeer), ctx, accountID, arg2, peersGroup, extraSettings)
+}
+
+// PeerDeleted mocks base method.
+func (m *MockIntegratedValidator) PeerDeleted(ctx context.Context, accountID, peerID string, extraSettings *types.ExtraSettings) error {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "PeerDeleted", ctx, accountID, peerID, extraSettings)
+	ret0, _ := ret[0].(error)
+	return ret0
+}
+
+// PeerDeleted indicates an expected call of PeerDeleted.
+func (mr *MockIntegratedValidatorMockRecorder) PeerDeleted(ctx, accountID, peerID, extraSettings any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "PeerDeleted", reflect.TypeOf((*MockIntegratedValidator)(nil).PeerDeleted), ctx, accountID, peerID, extraSettings)
+}
+
+// PreparePeer mocks base method.
+func (m *MockIntegratedValidator) PreparePeer(ctx context.Context, accountID string, p *peer.Peer, peersGroup []string, extraSettings *types.ExtraSettings, temporary bool) *peer.Peer {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "PreparePeer", ctx, accountID, p, peersGroup, extraSettings, temporary)
+	ret0, _ := ret[0].(*peer.Peer)
+	return ret0
+}
+
+// PreparePeer indicates an expected call of PreparePeer.
+func (mr *MockIntegratedValidatorMockRecorder) PreparePeer(ctx, accountID, p, peersGroup, extraSettings, temporary any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "PreparePeer", reflect.TypeOf((*MockIntegratedValidator)(nil).PreparePeer), ctx, accountID, p, peersGroup, extraSettings, temporary)
+}
+
+// SetPeerInvalidationListener mocks base method.
+func (m *MockIntegratedValidator) SetPeerInvalidationListener(fn func(string, []string)) {
+	m.ctrl.T.Helper()
+	m.ctrl.Call(m, "SetPeerInvalidationListener", fn)
+}
+
+// SetPeerInvalidationListener indicates an expected call of SetPeerInvalidationListener.
+func (mr *MockIntegratedValidatorMockRecorder) SetPeerInvalidationListener(fn any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetPeerInvalidationListener", reflect.TypeOf((*MockIntegratedValidator)(nil).SetPeerInvalidationListener), fn)
+}
+
+// Stop mocks base method.
+func (m *MockIntegratedValidator) Stop(ctx context.Context) {
+	m.ctrl.T.Helper()
+	m.ctrl.Call(m, "Stop", ctx)
+}
+
+// Stop indicates an expected call of Stop.
+func (mr *MockIntegratedValidatorMockRecorder) Stop(ctx any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Stop", reflect.TypeOf((*MockIntegratedValidator)(nil).Stop), ctx)
+}
+
+// ValidateExtraSettings mocks base method.
+func (m *MockIntegratedValidator) ValidateExtraSettings(ctx context.Context, newExtraSettings, oldExtraSettings *types.ExtraSettings, userID, accountID string) error {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "ValidateExtraSettings", ctx, newExtraSettings, oldExtraSettings, userID, accountID)
+	ret0, _ := ret[0].(error)
+	return ret0
+}
+
+// ValidateExtraSettings indicates an expected call of ValidateExtraSettings.
+func (mr *MockIntegratedValidatorMockRecorder) ValidateExtraSettings(ctx, newExtraSettings, oldExtraSettings, userID, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ValidateExtraSettings", reflect.TypeOf((*MockIntegratedValidator)(nil).ValidateExtraSettings), ctx, newExtraSettings, oldExtraSettings, userID, accountID)
+}
+
+// ValidateFlowResponse mocks base method.
+func (m *MockIntegratedValidator) ValidateFlowResponse(ctx context.Context, peerKey string, flowResponse *proto.PKCEAuthorizationFlow) *proto.PKCEAuthorizationFlow {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "ValidateFlowResponse", ctx, peerKey, flowResponse)
+	ret0, _ := ret[0].(*proto.PKCEAuthorizationFlow)
+	return ret0
+}
+
+// ValidateFlowResponse indicates an expected call of ValidateFlowResponse.
+func (mr *MockIntegratedValidatorMockRecorder) ValidateFlowResponse(ctx, peerKey, flowResponse any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ValidateFlowResponse", reflect.TypeOf((*MockIntegratedValidator)(nil).ValidateFlowResponse), ctx, peerKey, flowResponse)
+}
+
+// ValidatePeer mocks base method.
+func (m *MockIntegratedValidator) ValidatePeer(ctx context.Context, update, p *peer.Peer, userID, accountID, dnsDomain string, peersGroup []string, extraSettings *types.ExtraSettings) (*peer.Peer, bool, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "ValidatePeer", ctx, update, p, userID, accountID, dnsDomain, peersGroup, extraSettings)
+	ret0, _ := ret[0].(*peer.Peer)
+	ret1, _ := ret[1].(bool)
+	ret2, _ := ret[2].(error)
+	return ret0, ret1, ret2
+}
+
+// ValidatePeer indicates an expected call of ValidatePeer.
+func (mr *MockIntegratedValidatorMockRecorder) ValidatePeer(ctx, update, p, userID, accountID, dnsDomain, peersGroup, extraSettings any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ValidatePeer", reflect.TypeOf((*MockIntegratedValidator)(nil).ValidatePeer), ctx, update, p, userID, accountID, dnsDomain, peersGroup, extraSettings)
+}

management/server/integrations/integrated_validator/interface.go

@@ -5,16 +5,19 @@ import (
 
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/management/networkmap/nmdata"
 	"github.com/netbirdio/netbird/shared/management/proto"
 )
 
+//go:generate go tool mockgen -package integrated_validator -destination=integrated_validator_mock.go -source=./interface.go -build_flags=-mod=mod
+
 // IntegratedValidator interface exists to avoid the circle dependencies
 type IntegratedValidator interface {
 	ValidateExtraSettings(ctx context.Context, newExtraSettings *types.ExtraSettings, oldExtraSettings *types.ExtraSettings, userID string, accountID string) error
-	ValidatePeer(ctx context.Context, update *nbpeer.Peer, peer *nbpeer.Peer, userID string, accountID string, dnsDomain string, peersGroup []string, extraSettings *types.ExtraSettings) (*nbpeer.Peer, bool, error)
-	PreparePeer(ctx context.Context, accountID string, peer *nbpeer.Peer, peersGroup []string, extraSettings *types.ExtraSettings, temporary bool) *nbpeer.Peer
+	ValidatePeer(ctx context.Context, update *nbpeer.Peer, p *nbpeer.Peer, userID string, accountID string, dnsDomain string, peersGroup []string, extraSettings *types.ExtraSettings) (*nbpeer.Peer, bool, error)
+	PreparePeer(ctx context.Context, accountID string, p *nbpeer.Peer, peersGroup []string, extraSettings *types.ExtraSettings, temporary bool) *nbpeer.Peer
 	IsNotValidPeer(ctx context.Context, accountID string, peer *nbpeer.Peer, peersGroup []string, extraSettings *types.ExtraSettings) (bool, bool, error)
-	GetValidatedPeers(ctx context.Context, accountID string, groups []*types.Group, peers []*nbpeer.Peer, extraSettings *types.ExtraSettings) (map[string]struct{}, error)
+	GetValidatedPeers(ctx context.Context, accountID string, groups []*nmdata.Group, peers []*nmdata.Peer, extraSettings *types.ExtraSettings) (map[string]struct{}, error)
 	GetInvalidPeers(ctx context.Context, accountID string, extraSettings *types.ExtraSettings) (map[string]string, error)
 	PeerDeleted(ctx context.Context, accountID, peerID string, extraSettings *types.ExtraSettings) error
 	SetPeerInvalidationListener(fn func(accountID string, peerIDs []string))

management/server/integrations/integrated_validator/validator/validator.go

@@ -0,0 +1,63 @@
+package validator
+
+import (
+	"context"
+
+	cachestore "github.com/eko/gocache/lib/v4/store"
+
+	"github.com/netbirdio/netbird/management/internals/modules/peers"
+	"github.com/netbirdio/netbird/management/server/activity"
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/settings"
+	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/management/networkmap/nmdata"
+	"github.com/netbirdio/netbird/shared/management/proto"
+)
+
+type IntegratedValidatorImpl struct{}
+
+func NewIntegratedValidator(_ context.Context, _ peers.Manager, _ settings.Manager, _ activity.Store, _ cachestore.StoreInterface) (*IntegratedValidatorImpl, error) {
+	return &IntegratedValidatorImpl{}, nil
+}
+
+func (v *IntegratedValidatorImpl) ValidateExtraSettings(context.Context, *types.ExtraSettings, *types.ExtraSettings, string, string) error {
+	return nil
+}
+
+func (v *IntegratedValidatorImpl) ValidatePeer(_ context.Context, update *nbpeer.Peer, _ *nbpeer.Peer, _ string, _ string, _ string, _ []string, _ *types.ExtraSettings) (*nbpeer.Peer, bool, error) {
+	return update, false, nil
+}
+
+func (v *IntegratedValidatorImpl) PreparePeer(_ context.Context, _ string, peer *nbpeer.Peer, _ []string, _ *types.ExtraSettings, _ bool) *nbpeer.Peer {
+	return peer.Copy()
+}
+
+func (v *IntegratedValidatorImpl) IsNotValidPeer(_ context.Context, _ string, _ *nbpeer.Peer, _ []string, _ *types.ExtraSettings) (bool, bool, error) {
+	return false, false, nil
+}
+
+func (v *IntegratedValidatorImpl) GetValidatedPeers(_ context.Context, _ string, _ []*nmdata.Group, peers []*nmdata.Peer, _ *types.ExtraSettings) (map[string]struct{}, error) {
+	validatedPeers := make(map[string]struct{})
+	for _, p := range peers {
+		validatedPeers[p.ID] = struct{}{}
+	}
+	return validatedPeers, nil
+}
+
+func (v *IntegratedValidatorImpl) GetInvalidPeers(_ context.Context, _ string, _ *types.ExtraSettings) (map[string]string, error) {
+	return make(map[string]string), nil
+}
+
+func (v *IntegratedValidatorImpl) PeerDeleted(_ context.Context, _, _ string, _ *types.ExtraSettings) error {
+	return nil
+}
+
+func (v *IntegratedValidatorImpl) SetPeerInvalidationListener(_ func(accountID string, peerIDs []string)) {
+}
+
+func (v *IntegratedValidatorImpl) Stop(_ context.Context) {
+}
+
+func (v *IntegratedValidatorImpl) ValidateFlowResponse(_ context.Context, _ string, flowResponse *proto.PKCEAuthorizationFlow) *proto.PKCEAuthorizationFlow {
+	return flowResponse
+}

management/server/management_proto_test.go

@@ -12,7 +12,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/require"
 	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
@@ -376,7 +376,7 @@ func startManagementForTest(t *testing.T, testFile string, config *config.Config
 		return nil, nil, "", cleanup, err
 	}
 
-	networkMapController := controller.NewController(ctx, store, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), ephemeralMgr, config)
+	networkMapController := controller.NewController(ctx, store, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), ephemeralMgr, config, nil)
 	accountManager, err := BuildManager(ctx, nil, store, networkMapController, jobManager, nil, "",
 		eventStore, nil, false, MockIntegratedValidator{}, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
 
@@ -391,7 +391,7 @@ func startManagementForTest(t *testing.T, testFile string, config *config.Config
 		return nil, nil, "", cleanup, err
 	}
 
-	mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, MockIntegratedValidator{}, networkMapController, nil)
+	mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, MockIntegratedValidator{}, networkMapController, nil, nil)
 	if err != nil {
 		return nil, nil, "", cleanup, err
 	}
@@ -728,7 +728,7 @@ func Test_LoginPerformance(t *testing.T) {
 						}
 
 						login := func() error {
-							_, _, _, err = am.LoginPeer(context.Background(), peerLogin)
+							_, _, _, _, err = am.LoginPeer(context.Background(), peerLogin)
 							if err != nil {
 								t.Logf("failed to login peer: %v", err)
 								return err
@@ -746,7 +746,7 @@ func Test_LoginPerformance(t *testing.T) {
 
 						go func(peerLogin types.PeerLogin, counterStart *int32) {
 							defer wgPeer.Done()
-							_, _, _, err = am.LoginPeer(context.Background(), peerLogin)
+							_, _, _, _, err = am.LoginPeer(context.Background(), peerLogin)
 							if err != nil {
 								t.Logf("failed to login peer: %v", err)
 								return

management/server/management_test.go

@@ -10,10 +10,10 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
 	pb "github.com/golang/protobuf/proto" //nolint
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
+	"go.uber.org/mock/gomock"
 	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/credentials/insecure"
@@ -216,7 +216,7 @@ func startServer(
 
 	updateManager := update_channel.NewPeersUpdateManager(metrics)
 	requestBuffer := server.NewAccountRequestBuffer(ctx, str)
-	networkMapController := controller.NewController(ctx, str, metrics, updateManager, requestBuffer, server.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(str, peers.NewManager(str)), config)
+	networkMapController := controller.NewController(ctx, str, metrics, updateManager, requestBuffer, server.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(str, peers.NewManager(str)), config, nil)
 
 	accountManager, err := server.BuildManager(
 		context.Background(),
@@ -256,6 +256,7 @@ func startServer(
 		server.MockIntegratedValidator{},
 		networkMapController,
 		nil,
+		nil,
 	)
 	if err != nil {
 		t.Fatalf("failed creating management server: %v", err)

management/server/metrics/selfhosted.go

@@ -17,6 +17,7 @@ import (
 	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	log "github.com/sirupsen/logrus"
 
+	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 	nbversion "github.com/netbirdio/netbird/version"
 )
@@ -53,6 +54,8 @@ type DataSource interface {
 	GetAllAccounts(ctx context.Context) []*types.Account
 	GetStoreEngine() types.Engine
 	GetCustomDomainsCounts(ctx context.Context) (total int64, validated int64, err error)
+	GetProxyMetrics(ctx context.Context) (store.ProxyMetrics, error)
+	GetAgentNetworkMetrics(ctx context.Context) (store.AgentNetworkMetrics, error)
 }
 
 // ConnManager peer connection manager that holds state for current active connections
@@ -223,6 +226,12 @@ func (w *Worker) generateProperties(ctx context.Context) properties {
 		servicesAuthPassword      int
 		servicesAuthPin           int
 		servicesAuthOIDC          int
+		// Private-service signals — track adoption of NetBird-only mode
+		// (services backed by an embedded proxy peer + access groups).
+		servicesPrivate                int
+		servicesPrivateWithGroups      int
+		servicesPrivateAccessGroupsSum int
+		servicesWithDirectUpstream     int
 	)
 	start := time.Now()
 	metricsProperties := make(properties)
@@ -380,9 +389,38 @@ func (w *Worker) generateProperties(ctx context.Context) properties {
 			if service.Auth.BearerAuth != nil && service.Auth.BearerAuth.Enabled {
 				servicesAuthOIDC++
 			}
+
+			if service.Private {
+				servicesPrivate++
+				if len(service.AccessGroups) > 0 {
+					servicesPrivateWithGroups++
+				}
+				servicesPrivateAccessGroupsSum += len(service.AccessGroups)
+			}
+
+			for _, target := range service.Targets {
+				if target.Options.DirectUpstream {
+					servicesWithDirectUpstream++
+					break
+				}
+			}
 		}
 	}
 
+	// Proxy / BYOP cluster signals come from the proxies table aggregated
+	// across all accounts in a single store query; nil on FileStore.
+	proxyMetrics, err := w.dataSource.GetProxyMetrics(ctx)
+	if err != nil {
+		log.WithContext(ctx).Debugf("collect proxy metrics: %v", err)
+	}
+
+	// Agent-network adoption + usage, aggregated across all accounts in a few
+	// cheap queries; nil on FileStore.
+	agentNetworkMetrics, err := w.dataSource.GetAgentNetworkMetrics(ctx)
+	if err != nil {
+		log.WithContext(ctx).Debugf("collect agent network metrics: %v", err)
+	}
+
 	minActivePeerVersion, maxActivePeerVersion := getMinMaxVersion(peerActiveVersions)
 	metricsProperties["uptime"] = uptime
 	metricsProperties["accounts"] = accounts
@@ -430,8 +468,25 @@ func (w *Worker) generateProperties(ctx context.Context) properties {
 	metricsProperties["services_auth_password"] = servicesAuthPassword
 	metricsProperties["services_auth_pin"] = servicesAuthPin
 	metricsProperties["services_auth_oidc"] = servicesAuthOIDC
+	metricsProperties["services_private"] = servicesPrivate
+	metricsProperties["services_private_with_access_groups"] = servicesPrivateWithGroups
+	metricsProperties["services_private_access_groups_sum"] = servicesPrivateAccessGroupsSum
+	metricsProperties["services_with_direct_upstream"] = servicesWithDirectUpstream
+	metricsProperties["proxy_clusters"] = proxyMetrics.Clusters
+	metricsProperties["proxy_clusters_byop"] = proxyMetrics.ClustersBYOP
+	metricsProperties["proxy_clusters_private"] = proxyMetrics.ClustersPrivate
+	metricsProperties["proxies"] = proxyMetrics.Proxies
+	metricsProperties["proxies_connected"] = proxyMetrics.ProxiesConnected
 	metricsProperties["custom_domains"] = customDomains
 	metricsProperties["custom_domains_validated"] = customDomainsValidated
+	metricsProperties["agent_network_accounts"] = agentNetworkMetrics.Accounts
+	metricsProperties["agent_network_providers"] = agentNetworkMetrics.Providers
+	metricsProperties["agent_network_policies"] = agentNetworkMetrics.Policies
+	metricsProperties["agent_network_budget_rules"] = agentNetworkMetrics.BudgetRules
+	metricsProperties["agent_network_log_collection_enabled"] = agentNetworkMetrics.LogCollectionEnabled
+	metricsProperties["agent_network_input_tokens"] = agentNetworkMetrics.InputTokens
+	metricsProperties["agent_network_output_tokens"] = agentNetworkMetrics.OutputTokens
+	metricsProperties["agent_network_cost_usd"] = agentNetworkMetrics.CostUSD
 
 	for targetType, count := range servicesTargetType {
 		metricsProperties["services_target_type_"+string(targetType)] = count

management/server/metrics/selfhosted_test.go

@@ -12,6 +12,7 @@ import (
 	networkTypes "github.com/netbirdio/netbird/management/server/networks/types"
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/posture"
+	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/route"
 )
@@ -123,7 +124,7 @@ func (mockDatasource) GetAllAccounts(_ context.Context) []*types.Account {
 					Enabled: true,
 					Targets: []*rpservice.Target{
 						{TargetType: "peer"},
-						{TargetType: "host"},
+						{TargetType: "host", Options: rpservice.TargetOptions{DirectUpstream: true}},
 					},
 					Auth: rpservice.AuthConfig{
 						PasswordAuth: &rpservice.PasswordAuthConfig{Enabled: true},
@@ -141,6 +142,16 @@ func (mockDatasource) GetAllAccounts(_ context.Context) []*types.Account {
 					},
 					Meta: rpservice.Meta{Status: string(rpservice.StatusPending)},
 				},
+				{
+					ID:           "svc3-private",
+					Enabled:      true,
+					Private:      true,
+					AccessGroups: []string{"grp-eng", "grp-ops"},
+					Targets: []*rpservice.Target{
+						{TargetType: "cluster", Options: rpservice.TargetOptions{DirectUpstream: true}},
+					},
+					Meta: rpservice.Meta{Status: string(rpservice.StatusActive)},
+				},
 			},
 		},
 		{
@@ -254,6 +265,33 @@ func (mockDatasource) GetCustomDomainsCounts(_ context.Context) (int64, int64, e
 	return 3, 2, nil
 }
 
+// GetProxyMetrics returns canned proxy/cluster counts so the
+// generateProperties test can assert the BYOP signals end-to-end.
+func (mockDatasource) GetProxyMetrics(_ context.Context) (store.ProxyMetrics, error) {
+	return store.ProxyMetrics{
+		Clusters:         3,
+		ClustersBYOP:     1,
+		ClustersPrivate:  1,
+		Proxies:          4,
+		ProxiesConnected: 2,
+	}, nil
+}
+
+// GetAgentNetworkMetrics returns canned agent-network counts so the
+// generateProperties test can assert the adoption/usage signals end-to-end.
+func (mockDatasource) GetAgentNetworkMetrics(_ context.Context) (store.AgentNetworkMetrics, error) {
+	return store.AgentNetworkMetrics{
+		Accounts:             2,
+		Providers:            5,
+		Policies:             3,
+		BudgetRules:          1,
+		LogCollectionEnabled: 2,
+		InputTokens:          1000,
+		OutputTokens:         500,
+		CostUSD:              1.25,
+	}, nil
+}
+
 // TestGenerateProperties tests and validate the properties generation by using the mockDatasource for the Worker.generateProperties
 func TestGenerateProperties(t *testing.T) {
 	ds := mockDatasource{}
@@ -393,17 +431,17 @@ func TestGenerateProperties(t *testing.T) {
 		t.Errorf("expected 3 embedded_idp_count, got %v", properties["embedded_idp_count"])
 	}
 
-	if properties["services"] != 2 {
-		t.Errorf("expected 2 services, got %v", properties["services"])
+	if properties["services"] != 3 {
+		t.Errorf("expected 3 services, got %v", properties["services"])
 	}
-	if properties["services_enabled"] != 1 {
-		t.Errorf("expected 1 services_enabled, got %v", properties["services_enabled"])
+	if properties["services_enabled"] != 2 {
+		t.Errorf("expected 2 services_enabled, got %v", properties["services_enabled"])
 	}
-	if properties["services_targets"] != 3 {
-		t.Errorf("expected 3 services_targets, got %v", properties["services_targets"])
+	if properties["services_targets"] != 4 {
+		t.Errorf("expected 4 services_targets, got %v", properties["services_targets"])
 	}
-	if properties["services_status_active"] != 1 {
-		t.Errorf("expected 1 services_status_active, got %v", properties["services_status_active"])
+	if properties["services_status_active"] != 2 {
+		t.Errorf("expected 2 services_status_active, got %v", properties["services_status_active"])
 	}
 	if properties["services_status_pending"] != 1 {
 		t.Errorf("expected 1 services_status_pending, got %v", properties["services_status_pending"])
@@ -420,6 +458,9 @@ func TestGenerateProperties(t *testing.T) {
 	if properties["services_target_type_domain"] != 1 {
 		t.Errorf("expected 1 services_target_type_domain, got %v", properties["services_target_type_domain"])
 	}
+	if properties["services_target_type_cluster"] != 1 {
+		t.Errorf("expected 1 services_target_type_cluster, got %v", properties["services_target_type_cluster"])
+	}
 	if properties["services_auth_password"] != 1 {
 		t.Errorf("expected 1 services_auth_password, got %v", properties["services_auth_password"])
 	}
@@ -429,6 +470,33 @@ func TestGenerateProperties(t *testing.T) {
 	if properties["services_auth_pin"] != 0 {
 		t.Errorf("expected 0 services_auth_pin, got %v", properties["services_auth_pin"])
 	}
+	if properties["services_private"] != 1 {
+		t.Errorf("expected 1 services_private, got %v", properties["services_private"])
+	}
+	if properties["services_private_with_access_groups"] != 1 {
+		t.Errorf("expected 1 services_private_with_access_groups, got %v", properties["services_private_with_access_groups"])
+	}
+	if properties["services_private_access_groups_sum"] != 2 {
+		t.Errorf("expected 2 services_private_access_groups_sum, got %v", properties["services_private_access_groups_sum"])
+	}
+	if properties["services_with_direct_upstream"] != 2 {
+		t.Errorf("expected 2 services_with_direct_upstream, got %v", properties["services_with_direct_upstream"])
+	}
+	if properties["proxy_clusters"] != int64(3) {
+		t.Errorf("expected 3 proxy_clusters, got %v", properties["proxy_clusters"])
+	}
+	if properties["proxy_clusters_byop"] != int64(1) {
+		t.Errorf("expected 1 proxy_clusters_byop, got %v", properties["proxy_clusters_byop"])
+	}
+	if properties["proxy_clusters_private"] != int64(1) {
+		t.Errorf("expected 1 proxy_clusters_private, got %v", properties["proxy_clusters_private"])
+	}
+	if properties["proxies"] != int64(4) {
+		t.Errorf("expected 4 proxies, got %v", properties["proxies"])
+	}
+	if properties["proxies_connected"] != int64(2) {
+		t.Errorf("expected 2 proxies_connected, got %v", properties["proxies_connected"])
+	}
 	if properties["custom_domains"] != int64(3) {
 		t.Errorf("expected 3 custom_domains, got %v", properties["custom_domains"])
 	}

management/server/migration/migration.go

@@ -13,6 +13,7 @@ import (
 	"strings"
 	"unicode/utf8"
 
+	"github.com/rs/xid"
 	log "github.com/sirupsen/logrus"
 	"gorm.io/gorm"
 	"gorm.io/gorm/clause"
@@ -635,3 +636,128 @@ func RemoveDuplicatePeerKeys(ctx context.Context, db *gorm.DB) error {
 
 	return nil
 }
+
+func BackfillPublicIDs[T any](ctx context.Context, db *gorm.DB) error {
+	var model T
+
+	if !db.Migrator().HasTable(&model) {
+		log.WithContext(ctx).Debugf("Table for %T does not exist, no backfill needed", model)
+		return nil
+	}
+
+	stmt := &gorm.Statement{DB: db}
+	err := stmt.Parse(&model)
+	if err != nil {
+		return fmt.Errorf("parse model: %w", err)
+	}
+	tableName := stmt.Schema.Table
+
+	if err := db.Transaction(func(tx *gorm.DB) error {
+		if !tx.Migrator().HasColumn(&model, "public_id") {
+			log.WithContext(ctx).Infof("Column public_id does not exist in table %s, adding it", tableName)
+			if err := tx.Migrator().AddColumn(&model, "public_id"); err != nil {
+				return fmt.Errorf("add column public_id: %w", err)
+			}
+		}
+
+		var rows []map[string]any
+		if err := tx.Table(tableName).Select("id", "public_id").Where("public_id IS NULL").Or("public_id = ''").Find(&rows).Error; err != nil {
+			return fmt.Errorf("failed to find rows with empty public_id: %w", err)
+		}
+
+		if len(rows) == 0 {
+			log.WithContext(ctx).Infof("No rows with empty public_id found in table %s, no migration needed", tableName)
+			return nil
+		}
+
+		for _, row := range rows {
+			if err := tx.Table(tableName).Where("id = ?", row["id"]).Update("public_id", xid.New().String()).Error; err != nil {
+				return fmt.Errorf("failed to update row with id %v: %w", row["id"], err)
+			}
+		}
+		return nil
+	}); err != nil {
+		return err
+	}
+
+	log.WithContext(ctx).Infof("Backfill of empty public_id in table %s completed", tableName)
+	return nil
+}
+
+// FoldCostAggregatesIntoBuckets migrates a per-request cost table from the old
+// "stored aggregate" shape (cost_usd + cache_cost_usd columns) to the per-bucket
+// breakdown, where the total and cache portion are derived on read instead.
+//
+// The fold preserves both aggregates exactly for historical rows: the cache
+// total moves into cached_input_cost_usd and the remainder into
+// input_cost_usd, so a row's derived total and cache cost still match what it
+// reported before the upgrade. The finer split is genuinely unknown for those
+// rows — the old schema never recorded a read/write or input/output division —
+// so it is lumped rather than guessed; only rows written after the upgrade
+// carry a true four-way split.
+//
+// Dropping the columns before folding would zero every historical row's cost,
+// so the update runs first and the drop only happens once it succeeds. A table
+// with no cost_usd column has already been migrated (or was created fresh) and
+// is skipped.
+func FoldCostAggregatesIntoBuckets[T any](ctx context.Context, db *gorm.DB) error {
+	var model T
+
+	if !db.Migrator().HasTable(&model) {
+		log.WithContext(ctx).Debugf("table for %T does not exist, no cost-bucket migration needed", model)
+		return nil
+	}
+	if !db.Migrator().HasColumn(&model, "cost_usd") {
+		log.WithContext(ctx).Debugf("table for %T has no cost_usd column, cost buckets already migrated", model)
+		return nil
+	}
+
+	stmt := &gorm.Statement{DB: db}
+	if err := stmt.Parse(&model); err != nil {
+		return fmt.Errorf("parse model schema: %w", err)
+	}
+	tableName := stmt.Schema.Table
+
+	// COALESCE guards rows whose new columns were added as NULL by an earlier
+	// AutoMigrate run that predates the NOT NULL default.
+	hasCacheColumn := db.Migrator().HasColumn(&model, "cache_cost_usd")
+	cacheExpr := "0"
+	if hasCacheColumn {
+		cacheExpr = "COALESCE(cache_cost_usd, 0)"
+	}
+
+	if err := db.Transaction(func(tx *gorm.DB) error {
+		// Only touch rows that carry a legacy total and no breakdown yet, so
+		// the migration is idempotent and never overwrites a true split.
+		update := fmt.Sprintf(`UPDATE %s
+			SET input_cost_usd = COALESCE(cost_usd, 0) - %s,
+			    cached_input_cost_usd = %s,
+			    cache_creation_cost_usd = 0,
+			    output_cost_usd = 0
+			WHERE COALESCE(cost_usd, 0) <> 0
+			  AND COALESCE(input_cost_usd, 0) = 0
+			  AND COALESCE(cached_input_cost_usd, 0) = 0
+			  AND COALESCE(cache_creation_cost_usd, 0) = 0
+			  AND COALESCE(output_cost_usd, 0) = 0`, tableName, cacheExpr, cacheExpr)
+		res := tx.Exec(update)
+		if res.Error != nil {
+			return fmt.Errorf("fold legacy cost aggregates in %s: %w", tableName, res.Error)
+		}
+		log.WithContext(ctx).Infof("folded legacy cost aggregates into per-bucket columns for %d rows in table %s", res.RowsAffected, tableName)
+
+		if err := tx.Migrator().DropColumn(&model, "cost_usd"); err != nil {
+			return fmt.Errorf("drop cost_usd from %s: %w", tableName, err)
+		}
+		if hasCacheColumn {
+			if err := tx.Migrator().DropColumn(&model, "cache_cost_usd"); err != nil {
+				return fmt.Errorf("drop cache_cost_usd from %s: %w", tableName, err)
+			}
+		}
+		return nil
+	}); err != nil {
+		return err
+	}
+
+	log.WithContext(ctx).Infof("migration of stored cost aggregates to per-bucket columns in table %s completed", tableName)
+	return nil
+}

management/server/migration/migration_agentnetwork.go

@@ -0,0 +1,112 @@
+package migration
+
+import (
+	"context"
+	"fmt"
+
+	log "github.com/sirupsen/logrus"
+	"gorm.io/gorm"
+)
+
+// agentNetworkSettingsMigration is a local view of the agent_network_settings
+// table spanning both the legacy identity columns (cluster, subdomain) and
+// their replacement (domain, proxy_address), so the migrator can address all
+// four during the reshape without importing the current model.
+type agentNetworkSettingsMigration struct {
+	AccountID    string `gorm:"primaryKey"`
+	Cluster      string
+	Subdomain    string
+	Domain       string `gorm:"type:varchar(255)"`
+	ProxyAddress string `gorm:"type:varchar(255)"`
+}
+
+func (agentNetworkSettingsMigration) TableName() string { return "agent_network_settings" }
+
+// MigrateAgentNetworkSettingsToDomain reshapes agent_network_settings from the
+// legacy (cluster, subdomain) identity columns to (domain, proxy_address):
+// domain becomes `<subdomain>.<cluster>` — the endpoint hostname the old
+// columns derived — and proxy_address becomes the cluster address, preserving
+// which proxy serves the account. Runs before AutoMigrate, which then creates
+// the unique index on the freshly backfilled domain column.
+//
+// A legacy row missing either half cannot be given an endpoint; the old
+// bootstrap always wrote both, so such a row indicates corruption and the
+// migration fails loudly rather than leaving an empty domain to collide with
+// the unique index confusingly.
+//
+// The transaction is real only on sqlite and postgres, where DDL is
+// transactional. MySQL implicitly commits around every ALTER TABLE, so there
+// each step stands alone; what makes an interrupted run resumable on MySQL is
+// that every step is guarded by the schema state it changes — the entry check
+// fires while either legacy column remains, the adds skip existing columns,
+// the backfill and its loud-failure check run only while the legacy cluster
+// column exists (they provably completed before any drop), and each drop
+// skips what is already gone.
+func MigrateAgentNetworkSettingsToDomain(ctx context.Context, db *gorm.DB) error {
+	model := &agentNetworkSettingsMigration{}
+	migrator := db.Migrator()
+
+	if !migrator.HasTable(model) {
+		return nil
+	}
+	hasCluster := migrator.HasColumn(model, "cluster")
+	if !hasCluster && !migrator.HasColumn(model, "subdomain") {
+		// Fresh schema or already migrated — nothing to reshape.
+		return nil
+	}
+
+	return db.Transaction(func(tx *gorm.DB) error {
+		txMigrator := tx.Migrator()
+		for _, field := range []string{"Domain", "ProxyAddress"} {
+			if !txMigrator.HasColumn(model, field) {
+				if err := txMigrator.AddColumn(model, field); err != nil {
+					return fmt.Errorf("add %s column to agent_network_settings: %w", field, err)
+				}
+			}
+		}
+
+		if hasCluster {
+			concat := "subdomain || '.' || cluster"
+			if tx.Name() == "mysql" {
+				concat = "CONCAT(subdomain, '.', cluster)"
+			}
+			res := tx.Exec(fmt.Sprintf(
+				"UPDATE agent_network_settings SET domain = %s, proxy_address = cluster WHERE (domain IS NULL OR domain = '') AND cluster <> '' AND subdomain <> ''",
+				concat,
+			))
+			if res.Error != nil {
+				return fmt.Errorf("backfill agent_network_settings domain: %w", res.Error)
+			}
+
+			var unmigratable int64
+			if err := tx.Model(model).Where("domain IS NULL OR domain = ''").Count(&unmigratable).Error; err != nil {
+				return fmt.Errorf("count unmigratable agent_network_settings rows: %w", err)
+			}
+			if unmigratable > 0 {
+				return fmt.Errorf(
+					"%d agent_network_settings row(s) have no cluster/subdomain to derive an endpoint from; resolve them manually before upgrading",
+					unmigratable,
+				)
+			}
+
+			if res.RowsAffected > 0 {
+				log.WithContext(ctx).Infof("migrated %d agent_network_settings row(s) to domain/proxy_address", res.RowsAffected)
+			}
+		}
+
+		if txMigrator.HasIndex(model, "idx_agent_network_settings_cluster_subdomain") {
+			if err := txMigrator.DropIndex(model, "idx_agent_network_settings_cluster_subdomain"); err != nil {
+				return fmt.Errorf("drop legacy agent_network_settings index: %w", err)
+			}
+		}
+		for _, field := range []string{"Cluster", "Subdomain"} {
+			if txMigrator.HasColumn(model, field) {
+				if err := txMigrator.DropColumn(model, field); err != nil {
+					return fmt.Errorf("drop legacy agent_network_settings column %s: %w", field, err)
+				}
+			}
+		}
+
+		return nil
+	})
+}

management/server/migration/migration_test.go

@@ -16,6 +16,7 @@ import (
 	"gorm.io/driver/sqlite"
 	"gorm.io/gorm"
 
+	agentNetworkTypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
 	"github.com/netbirdio/netbird/management/server/migration"
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/testutil"
@@ -198,7 +199,11 @@ func TestMigrateNetIPFieldFromBlobToJSON_WithJSONData(t *testing.T) {
 	require.NoError(t, err, "Failed to insert account")
 
 	account.PeersG = []nbpeer.Peer{
-		{AccountID: "1234", Location: nbpeer.Location{ConnectionIP: net.IP{10, 0, 0, 1}}},
+		{
+			AccountID: "1234",
+			Location:  nbpeer.Location{ConnectionIP: net.IP{10, 0, 0, 1}},
+			Status:    &nbpeer.PeerStatus{LastSeen: time.Now()},
+		},
 	}
 
 	err = db.Save(account).Error
@@ -635,3 +640,221 @@ func TestCleanupOrphanedResources_SkipsWhenForeignKeyExists(t *testing.T) {
 	db.Model(&testChildWithFK{}).Count(&count)
 	assert.Equal(t, int64(2), count, "Both rows should survive — migration must skip when FK constraint exists")
 }
+
+// legacyCostRow is the pre-breakdown shape of the usage table: cost was stored
+// as a total plus a cache portion, with no per-bucket columns. Used to build a
+// realistic pre-upgrade table for the fold migration to run against.
+type legacyCostRow struct {
+	ID           string `gorm:"primaryKey"`
+	AccountID    string
+	Model        string
+	CostUSD      float64
+	CacheCostUSD float64
+}
+
+func (legacyCostRow) TableName() string { return "agent_network_request_usage" }
+
+// TestFoldCostAggregatesIntoBuckets_PreservesHistoricalCost covers the upgrade
+// path: a table written under the old schema must come out with its per-row
+// total and cache cost unchanged, because dropping cost_usd without folding it
+// forward would silently zero every historical row's spend.
+func TestFoldCostAggregatesIntoBuckets_PreservesHistoricalCost(t *testing.T) {
+	ctx := context.Background()
+	db := setupDatabase(t)
+	// setupDatabase hands back a process-shared database, so start from a clean
+	// table rather than inheriting rows from another test.
+	require.NoError(t, db.Migrator().DropTable(&agentNetworkTypes.AgentNetworkUsage{}))
+
+	require.NoError(t, db.AutoMigrate(&legacyCostRow{}), "legacy table must be created")
+	require.NoError(t, db.Create(&legacyCostRow{
+		ID: "u1", AccountID: "acct-1", Model: "claude-sonnet-4-6", CostUSD: 0.0123, CacheCostUSD: 0.0029,
+	}).Error)
+	require.NoError(t, db.Create(&legacyCostRow{
+		ID: "u2", AccountID: "acct-1", Model: "gpt-4o", CostUSD: 0.5, CacheCostUSD: 0,
+	}).Error)
+	// A zero-cost row (denied / unpriced request) must stay zero, not be touched.
+	require.NoError(t, db.Create(&legacyCostRow{ID: "u3", AccountID: "acct-1", Model: "gw/unpriced"}).Error)
+
+	// AutoMigrate adds the per-bucket columns alongside the legacy ones, exactly
+	// as a real upgrade does before the post-auto migrations run.
+	require.NoError(t, db.AutoMigrate(&agentNetworkTypes.AgentNetworkUsage{}), "new columns must be added")
+
+	require.NoError(t, migration.FoldCostAggregatesIntoBuckets[agentNetworkTypes.AgentNetworkUsage](ctx, db))
+
+	assert.False(t, db.Migrator().HasColumn(&agentNetworkTypes.AgentNetworkUsage{}, "cost_usd"),
+		"legacy cost_usd column must be dropped once folded")
+	assert.False(t, db.Migrator().HasColumn(&agentNetworkTypes.AgentNetworkUsage{}, "cache_cost_usd"),
+		"legacy cache_cost_usd column must be dropped once folded")
+
+	var rows []*agentNetworkTypes.AgentNetworkUsage
+	require.NoError(t, db.Order("id").Find(&rows).Error)
+	require.Len(t, rows, 3)
+
+	// u1: total and cache portion both preserved; the read/write and
+	// input/output splits are unknowable for a legacy row, so the cache total
+	// lands on cached_input and the remainder on input.
+	assert.InDelta(t, 0.0123, rows[0].TotalCostUSD(), 1e-9, "historical total must survive the fold")
+	assert.InDelta(t, 0.0029, rows[0].CacheCostUSD(), 1e-9, "historical cache cost must survive the fold")
+	assert.InDelta(t, 0.0094, rows[0].InputCostUSD, 1e-9, "non-cache remainder lands on input")
+	assert.InDelta(t, 0.0029, rows[0].CachedInputCostUSD, 1e-9, "legacy cache total lands on cached input")
+	assert.Zero(t, rows[0].CacheCreationCostUSD, "legacy rows carry no read/write split to recover")
+	assert.Zero(t, rows[0].OutputCostUSD, "legacy rows carry no input/output split to recover")
+
+	// u2: no cache spend — the whole total is the non-cache remainder.
+	assert.InDelta(t, 0.5, rows[1].TotalCostUSD(), 1e-9, "cache-free historical total must survive")
+	assert.Zero(t, rows[1].CacheCostUSD(), "a cache-free row must stay cache-free")
+
+	// u3: zero stays zero rather than being rewritten.
+	assert.Zero(t, rows[2].TotalCostUSD(), "an unpriced row must remain unpriced")
+}
+
+// TestFoldCostAggregatesIntoBuckets_SkipsAlreadyMigrated proves the migration is
+// safe to re-run: with no legacy column present it is a no-op that leaves a
+// true four-way split untouched.
+func TestFoldCostAggregatesIntoBuckets_SkipsAlreadyMigrated(t *testing.T) {
+	ctx := context.Background()
+	db := setupDatabase(t)
+	require.NoError(t, db.Migrator().DropTable(&agentNetworkTypes.AgentNetworkUsage{}))
+
+	require.NoError(t, db.AutoMigrate(&agentNetworkTypes.AgentNetworkUsage{}))
+	// Timestamp must be set explicitly: a zero time.Time serialises as
+	// '0000-00-00 00:00:00', which MySQL rejects under strict mode.
+	require.NoError(t, db.Create(&agentNetworkTypes.AgentNetworkUsage{
+		ID: "u1", AccountID: "acct-1", Model: "claude-sonnet-4-6",
+		Timestamp:    time.Date(2026, 5, 5, 9, 0, 0, 0, time.UTC),
+		InputCostUSD: 0.001, CachedInputCostUSD: 0.002, CacheCreationCostUSD: 0.003, OutputCostUSD: 0.004,
+	}).Error)
+
+	require.NoError(t, migration.FoldCostAggregatesIntoBuckets[agentNetworkTypes.AgentNetworkUsage](ctx, db),
+		"running against an already-migrated table must be a no-op, not an error")
+
+	var row agentNetworkTypes.AgentNetworkUsage
+	require.NoError(t, db.First(&row, "id = ?", "u1").Error)
+	assert.InDelta(t, 0.001, row.InputCostUSD, 1e-9, "a true split must not be rewritten")
+	assert.InDelta(t, 0.002, row.CachedInputCostUSD, 1e-9)
+	assert.InDelta(t, 0.003, row.CacheCreationCostUSD, 1e-9)
+	assert.InDelta(t, 0.004, row.OutputCostUSD, 1e-9)
+	assert.InDelta(t, 0.01, row.TotalCostUSD(), 1e-9, "derived total sums the four buckets")
+}
+
+// legacyAgentNetworkSettings is the pre-reshape schema: identity carried as
+// (cluster, subdomain) instead of (domain, proxy_address).
+type legacyAgentNetworkSettings struct {
+	AccountID           string `gorm:"primaryKey"`
+	Cluster             string
+	Subdomain           string
+	EnableLogCollection bool
+}
+
+func (legacyAgentNetworkSettings) TableName() string { return "agent_network_settings" }
+
+// TestMigrateAgentNetworkSettingsToDomain_BackfillsAndDropsLegacyColumns pins
+// the reshape: domain becomes `<subdomain>.<cluster>`, proxy_address becomes
+// the cluster, the legacy columns are dropped, and non-identity fields ride
+// through untouched.
+func TestMigrateAgentNetworkSettingsToDomain_BackfillsAndDropsLegacyColumns(t *testing.T) {
+	ctx := context.Background()
+	db := setupDatabase(t)
+	require.NoError(t, db.Migrator().DropTable(&legacyAgentNetworkSettings{}))
+	require.NoError(t, db.AutoMigrate(&legacyAgentNetworkSettings{}))
+	require.NoError(t, db.Create(&legacyAgentNetworkSettings{
+		AccountID: "acct-1", Cluster: "eu.proxy.netbird.io", Subdomain: "violet", EnableLogCollection: true,
+	}).Error)
+	require.NoError(t, db.Create(&legacyAgentNetworkSettings{
+		AccountID: "acct-2", Cluster: "us.proxy.netbird.io", Subdomain: "violet",
+	}).Error)
+
+	require.NoError(t, migration.MigrateAgentNetworkSettingsToDomain(ctx, db))
+	require.NoError(t, db.AutoMigrate(&agentNetworkTypes.Settings{}),
+		"AutoMigrate must create the domain unique index over the backfilled values")
+
+	var one, two agentNetworkTypes.Settings
+	require.NoError(t, db.First(&one, "account_id = ?", "acct-1").Error)
+	assert.Equal(t, "violet.eu.proxy.netbird.io", one.Domain, "domain must combine subdomain and cluster")
+	assert.Equal(t, "eu.proxy.netbird.io", one.ProxyAddress, "proxy address must carry the cluster")
+	assert.True(t, one.EnableLogCollection, "non-identity fields must ride through")
+	require.NoError(t, db.First(&two, "account_id = ?", "acct-2").Error)
+	assert.Equal(t, "violet.us.proxy.netbird.io", two.Domain,
+		"duplicate labels on different clusters are distinct hostnames and must both survive")
+
+	migrator := db.Migrator()
+	assert.False(t, migrator.HasColumn(&legacyAgentNetworkSettings{}, "cluster"), "legacy cluster column must be dropped")
+	assert.False(t, migrator.HasColumn(&legacyAgentNetworkSettings{}, "subdomain"), "legacy subdomain column must be dropped")
+}
+
+// TestMigrateAgentNetworkSettingsToDomain_SkipsAlreadyMigrated proves the
+// migration is safe to re-run: with no legacy column present it is a no-op.
+func TestMigrateAgentNetworkSettingsToDomain_SkipsAlreadyMigrated(t *testing.T) {
+	ctx := context.Background()
+	db := setupDatabase(t)
+	require.NoError(t, db.Migrator().DropTable(&agentNetworkTypes.Settings{}))
+	require.NoError(t, db.AutoMigrate(&agentNetworkTypes.Settings{}))
+	require.NoError(t, db.Create(&agentNetworkTypes.Settings{
+		AccountID: "acct-1", Domain: "gw.example.com", ProxyAddress: "gw.example.com",
+	}).Error)
+
+	require.NoError(t, migration.MigrateAgentNetworkSettingsToDomain(ctx, db),
+		"running against an already-migrated table must be a no-op, not an error")
+
+	var row agentNetworkTypes.Settings
+	require.NoError(t, db.First(&row, "account_id = ?", "acct-1").Error)
+	assert.Equal(t, "gw.example.com", row.Domain, "migrated rows must be untouched")
+}
+
+// TestMigrateAgentNetworkSettingsToDomain_FailsOnUnmigratableRow pins the
+// loud-failure contract: a legacy row missing its identity halves cannot be
+// given an endpoint, and silently leaving an empty domain would collide with
+// the unique index confusingly later.
+func TestMigrateAgentNetworkSettingsToDomain_FailsOnUnmigratableRow(t *testing.T) {
+	ctx := context.Background()
+	db := setupDatabase(t)
+	require.NoError(t, db.Migrator().DropTable(&legacyAgentNetworkSettings{}))
+	require.NoError(t, db.AutoMigrate(&legacyAgentNetworkSettings{}))
+	require.NoError(t, db.Create(&legacyAgentNetworkSettings{
+		AccountID: "acct-broken", Cluster: "", Subdomain: "",
+	}).Error)
+
+	err := migration.MigrateAgentNetworkSettingsToDomain(ctx, db)
+	require.Error(t, err, "a row with no identity to derive an endpoint from must fail the migration")
+	assert.Contains(t, err.Error(), "resolve them manually", "the error must tell the operator what to do")
+}
+
+// partialAgentNetworkSettings models the one non-atomic state a MySQL run can
+// be interrupted in: DDL auto-commits there, so a crash between the two legacy
+// column drops leaves subdomain behind while cluster (and the completed
+// backfill) are already committed.
+type partialAgentNetworkSettings struct {
+	AccountID    string `gorm:"primaryKey"`
+	Subdomain    string
+	Domain       string `gorm:"type:varchar(255)"`
+	ProxyAddress string `gorm:"type:varchar(255)"`
+}
+
+func (partialAgentNetworkSettings) TableName() string { return "agent_network_settings" }
+
+// TestMigrateAgentNetworkSettingsToDomain_ResumesAfterPartialDrop pins MySQL
+// resumability: a rerun over the interrupted state must remove the leftover
+// subdomain column without re-running the backfill (the cluster column that
+// feeds it is gone) and without touching the migrated values.
+func TestMigrateAgentNetworkSettingsToDomain_ResumesAfterPartialDrop(t *testing.T) {
+	ctx := context.Background()
+	db := setupDatabase(t)
+	require.NoError(t, db.Migrator().DropTable(&partialAgentNetworkSettings{}))
+	require.NoError(t, db.AutoMigrate(&partialAgentNetworkSettings{}))
+	require.NoError(t, db.Create(&partialAgentNetworkSettings{
+		AccountID: "acct-1", Subdomain: "violet",
+		Domain: "violet.eu.proxy.netbird.io", ProxyAddress: "eu.proxy.netbird.io",
+	}).Error)
+
+	require.NoError(t, migration.MigrateAgentNetworkSettingsToDomain(ctx, db),
+		"a rerun over a partially-dropped schema must resume, not error")
+
+	migrator := db.Migrator()
+	assert.False(t, migrator.HasColumn(&partialAgentNetworkSettings{}, "subdomain"),
+		"the leftover legacy column must be dropped on resume")
+
+	var row agentNetworkTypes.Settings
+	require.NoError(t, db.First(&row, "account_id = ?", "acct-1").Error)
+	assert.Equal(t, "violet.eu.proxy.netbird.io", row.Domain, "migrated values must be untouched")
+	assert.Equal(t, "eu.proxy.netbird.io", row.ProxyAddress, "migrated values must be untouched")
+}

management/server/mock_server/account_mock.go

@@ -15,6 +15,7 @@ import (
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	"github.com/netbirdio/netbird/management/server/idp"
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/posture"
@@ -23,6 +24,7 @@ import (
 	"github.com/netbirdio/netbird/management/server/users"
 	"github.com/netbirdio/netbird/route"
 	"github.com/netbirdio/netbird/shared/management/domain"
+	"github.com/netbirdio/netbird/shared/management/networkmap/nmdata"
 )
 
 var _ account.Manager = (*MockAccountManager)(nil)
@@ -38,12 +40,13 @@ type MockAccountManager struct {
 	GetUserFromUserAuthFunc               func(ctx context.Context, userAuth auth.UserAuth) (*types.User, error)
 	ListUsersFunc                         func(ctx context.Context, accountID string) ([]*types.User, error)
 	GetPeersFunc                          func(ctx context.Context, accountID, userID, nameFilter, ipFilter string, all bool) ([]*nbpeer.Peer, error)
-	MarkPeerConnectedFunc                 func(ctx context.Context, peerKey string, connected bool, realIP net.IP, syncTime time.Time) error
-	SyncAndMarkPeerFunc                   func(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error)
+	MarkPeerConnectedFunc                 func(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error
+	MarkPeerDisconnectedFunc              func(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64) error
+	SyncAndMarkPeerFunc                   func(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*nmdata.PostureChecks, int64, error)
 	DeletePeerFunc                        func(ctx context.Context, accountID, peerKey, userID string) error
 	GetNetworkMapFunc                     func(ctx context.Context, peerKey string) (*types.NetworkMap, error)
 	GetPeerNetworkFunc                    func(ctx context.Context, peerKey string) (*types.Network, error)
-	AddPeerFunc                           func(ctx context.Context, accountID string, setupKey string, userId string, peer *nbpeer.Peer, temporary bool) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, error)
+	AddPeerFunc                           func(ctx context.Context, accountID string, setupKey string, userId string, peer *nbpeer.Peer, temporary bool) (*nbpeer.Peer, *types.Network, []*nmdata.PostureChecks, bool, error)
 	GetGroupFunc                          func(ctx context.Context, accountID, groupID, userID string) (*types.Group, error)
 	GetAllGroupsFunc                      func(ctx context.Context, accountID, userID string) ([]*types.Group, error)
 	GetGroupByNameFunc                    func(ctx context.Context, groupName, accountID string) (*types.Group, error)
@@ -63,6 +66,7 @@ type MockAccountManager struct {
 	UpdatePeerMetaFunc                    func(ctx context.Context, peerID string, meta nbpeer.PeerSystemMeta) error
 	UpdatePeerFunc                        func(ctx context.Context, accountID, userID string, peer *nbpeer.Peer) (*nbpeer.Peer, error)
 	UpdatePeerIPFunc                      func(ctx context.Context, accountID, userID, peerID string, newIP netip.Addr) error
+	UpdatePeerIPv6Func                    func(ctx context.Context, accountID, userID, peerID string, newIPv6 netip.Addr) error
 	CreateRouteFunc                       func(ctx context.Context, accountID string, prefix netip.Prefix, networkType route.NetworkType, domains domain.List, peer string, peerGroups []string, description string, netID route.NetID, masquerade bool, metric int, groups, accessControlGroupIDs []string, enabled bool, userID string, keepRoute bool, isSelected bool) (*route.Route, error)
 	GetRouteFunc                          func(ctx context.Context, accountID string, routeID route.ID, userID string) (*route.Route, error)
 	SaveRouteFunc                         func(ctx context.Context, accountID string, userID string, route *route.Route) error
@@ -95,8 +99,9 @@ type MockAccountManager struct {
 	SaveDNSSettingsFunc                   func(ctx context.Context, accountID, userID string, dnsSettingsToSave *types.DNSSettings) error
 	GetPeerFunc                           func(ctx context.Context, accountID, peerID, userID string) (*nbpeer.Peer, error)
 	UpdateAccountSettingsFunc             func(ctx context.Context, accountID, userID string, newSettings *types.Settings) (*types.Settings, error)
-	LoginPeerFunc                         func(ctx context.Context, login types.PeerLogin) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, error)
-	SyncPeerFunc                          func(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error)
+	LoginPeerFunc                         func(ctx context.Context, login types.PeerLogin) (*nbpeer.Peer, *types.Network, []*nmdata.PostureChecks, bool, error)
+	ExtendPeerSessionFunc                 func(ctx context.Context, peerPubKey, userID string) (time.Time, error)
+	SyncPeerFunc                          func(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*nmdata.PostureChecks, int64, error)
 	InviteUserFunc                        func(ctx context.Context, accountID string, initiatorUserID string, targetUserEmail string) error
 	ApproveUserFunc                       func(ctx context.Context, accountID, initiatorUserID, targetUserID string) (*types.UserInfo, error)
 	RejectUserFunc                        func(ctx context.Context, accountID, initiatorUserID, targetUserID string) error
@@ -110,7 +115,7 @@ type MockAccountManager struct {
 	GetIdpManagerFunc                     func() idp.Manager
 	UpdateIntegratedValidatorFunc         func(ctx context.Context, accountID, userID, validator string, groups []string) error
 	GroupValidationFunc                   func(ctx context.Context, accountId string, groups []string) (bool, error)
-	SyncPeerMetaFunc                      func(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error
+	SyncPeerMetaFunc                      func(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error
 	FindExistingPostureCheckFunc          func(accountID string, checks *posture.ChecksDefinition) (*posture.Checks, error)
 	GetAccountIDForPeerKeyFunc            func(ctx context.Context, peerKey string) (string, error)
 	GetAccountByIDFunc                    func(ctx context.Context, accountID string, userID string) (*types.Account, error)
@@ -128,8 +133,9 @@ type MockAccountManager struct {
 	GetOrCreateAccountByPrivateDomainFunc func(ctx context.Context, initiatorId, domain string) (*types.Account, bool, error)
 
 	AllowSyncFunc                  func(string, uint64) bool
-	UpdateAccountPeersFunc         func(ctx context.Context, accountID string)
-	BufferUpdateAccountPeersFunc   func(ctx context.Context, accountID string)
+	UpdateAccountPeersFunc         func(ctx context.Context, accountID string, reason types.UpdateReason)
+	ExpandAndUpdateAffectedFunc    func(ctx context.Context, accountID string, snap *affectedpeers.Snapshot, change affectedpeers.Change)
+	BufferUpdateAccountPeersFunc   func(ctx context.Context, accountID string, reason types.UpdateReason)
 	RecalculateNetworkMapCacheFunc func(ctx context.Context, accountId string) error
 
 	GetIdentityProviderFunc    func(ctx context.Context, accountID, idpID, userID string) (*types.IdentityProvider, error)
@@ -200,15 +206,21 @@ func (am *MockAccountManager) UpdateGroups(ctx context.Context, accountID, userI
 	return status.Errorf(codes.Unimplemented, "method UpdateGroups is not implemented")
 }
 
-func (am *MockAccountManager) UpdateAccountPeers(ctx context.Context, accountID string) {
+func (am *MockAccountManager) UpdateAccountPeers(ctx context.Context, accountID string, reason types.UpdateReason) {
 	if am.UpdateAccountPeersFunc != nil {
-		am.UpdateAccountPeersFunc(ctx, accountID)
+		am.UpdateAccountPeersFunc(ctx, accountID, reason)
 	}
 }
 
-func (am *MockAccountManager) BufferUpdateAccountPeers(ctx context.Context, accountID string) {
+func (am *MockAccountManager) ExpandAndUpdateAffected(ctx context.Context, accountID string, snap *affectedpeers.Snapshot, change affectedpeers.Change) {
+	if am.ExpandAndUpdateAffectedFunc != nil {
+		am.ExpandAndUpdateAffectedFunc(ctx, accountID, snap, change)
+	}
+}
+
+func (am *MockAccountManager) BufferUpdateAccountPeers(ctx context.Context, accountID string, reason types.UpdateReason) {
 	if am.BufferUpdateAccountPeersFunc != nil {
-		am.BufferUpdateAccountPeersFunc(ctx, accountID)
+		am.BufferUpdateAccountPeersFunc(ctx, accountID, reason)
 	}
 }
 
@@ -219,14 +231,21 @@ func (am *MockAccountManager) DeleteSetupKey(ctx context.Context, accountID, use
 	return status.Errorf(codes.Unimplemented, "method DeleteSetupKey is not implemented")
 }
 
-func (am *MockAccountManager) SyncAndMarkPeer(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error) {
+func (am *MockAccountManager) SyncAndMarkPeer(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*nmdata.PostureChecks, int64, error) {
 	if am.SyncAndMarkPeerFunc != nil {
 		return am.SyncAndMarkPeerFunc(ctx, accountID, peerPubKey, meta, realIP, syncTime)
 	}
 	return nil, nil, nil, 0, status.Errorf(codes.Unimplemented, "method MarkPeerConnected is not implemented")
 }
 
-func (am *MockAccountManager) OnPeerDisconnected(_ context.Context, accountID string, peerPubKey string, streamStartTime time.Time) error {
+func (am *MockAccountManager) OnPeerDisconnected(ctx context.Context, accountID string, peerPubKey string, streamStartTime time.Time) error {
+	// Mirror DefaultAccountManager.OnPeerDisconnected: drive the fencing
+	// hook so tests that inject MarkPeerDisconnectedFunc actually observe
+	// disconnect events. Falls through to nil when no hook is set, which
+	// is the original behaviour.
+	if am.MarkPeerDisconnectedFunc != nil {
+		return am.MarkPeerDisconnectedFunc(ctx, peerPubKey, accountID, streamStartTime.UnixNano())
+	}
 	return nil
 }
 
@@ -327,13 +346,21 @@ func (am *MockAccountManager) GetAccountIDByUserID(ctx context.Context, userAuth
 }
 
 // MarkPeerConnected mock implementation of MarkPeerConnected from server.AccountManager interface
-func (am *MockAccountManager) MarkPeerConnected(ctx context.Context, peerKey string, connected bool, realIP net.IP, accountID string, syncTime time.Time) error {
+func (am *MockAccountManager) MarkPeerConnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
 	if am.MarkPeerConnectedFunc != nil {
-		return am.MarkPeerConnectedFunc(ctx, peerKey, connected, realIP, syncTime)
+		return am.MarkPeerConnectedFunc(ctx, peerKey, accountID, sessionStartedAt, nmap)
 	}
 	return status.Errorf(codes.Unimplemented, "method MarkPeerConnected is not implemented")
 }
 
+// MarkPeerDisconnected mock implementation of MarkPeerDisconnected from server.AccountManager interface
+func (am *MockAccountManager) MarkPeerDisconnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64) error {
+	if am.MarkPeerDisconnectedFunc != nil {
+		return am.MarkPeerDisconnectedFunc(ctx, peerKey, accountID, sessionStartedAt)
+	}
+	return status.Errorf(codes.Unimplemented, "method MarkPeerDisconnected is not implemented")
+}
+
 // DeleteAccount mock implementation of DeleteAccount from server.AccountManager interface
 func (am *MockAccountManager) DeleteAccount(ctx context.Context, accountID, userID string) error {
 	if am.DeleteAccountFunc != nil {
@@ -398,11 +425,11 @@ func (am *MockAccountManager) AddPeer(
 	userId string,
 	peer *nbpeer.Peer,
 	temporary bool,
-) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, error) {
+) (*nbpeer.Peer, *types.Network, []*nmdata.PostureChecks, bool, error) {
 	if am.AddPeerFunc != nil {
 		return am.AddPeerFunc(ctx, accountID, setupKey, userId, peer, temporary)
 	}
-	return nil, nil, nil, status.Errorf(codes.Unimplemented, "method AddPeer is not implemented")
+	return nil, nil, nil, false, status.Errorf(codes.Unimplemented, "method AddPeer is not implemented")
 }
 
 // GetGroupByName mock implementation of GetGroupByName from server.AccountManager interface
@@ -539,6 +566,13 @@ func (am *MockAccountManager) UpdatePeerIP(ctx context.Context, accountID, userI
 	return status.Errorf(codes.Unimplemented, "method UpdatePeerIP is not implemented")
 }
 
+func (am *MockAccountManager) UpdatePeerIPv6(ctx context.Context, accountID, userID, peerID string, newIPv6 netip.Addr) error {
+	if am.UpdatePeerIPv6Func != nil {
+		return am.UpdatePeerIPv6Func(ctx, accountID, userID, peerID, newIPv6)
+	}
+	return status.Errorf(codes.Unimplemented, "method UpdatePeerIPv6 is not implemented")
+}
+
 // CreateRoute mock implementation of CreateRoute from server.AccountManager interface
 func (am *MockAccountManager) CreateRoute(ctx context.Context, accountID string, prefix netip.Prefix, networkType route.NetworkType, domains domain.List, peerID string, peerGroupIDs []string, description string, netID route.NetID, masquerade bool, metric int, groups, accessControlGroupID []string, enabled bool, userID string, keepRoute bool, isSelected bool) (*route.Route, error) {
 	if am.CreateRouteFunc != nil {
@@ -829,15 +863,23 @@ func (am *MockAccountManager) UpdateAccountSettings(ctx context.Context, account
 }
 
 // LoginPeer mocks LoginPeer of the AccountManager interface
-func (am *MockAccountManager) LoginPeer(ctx context.Context, login types.PeerLogin) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, error) {
+func (am *MockAccountManager) LoginPeer(ctx context.Context, login types.PeerLogin) (*nbpeer.Peer, *types.Network, []*nmdata.PostureChecks, bool, error) {
 	if am.LoginPeerFunc != nil {
 		return am.LoginPeerFunc(ctx, login)
 	}
-	return nil, nil, nil, status.Errorf(codes.Unimplemented, "method LoginPeer is not implemented")
+	return nil, nil, nil, false, status.Errorf(codes.Unimplemented, "method LoginPeer is not implemented")
+}
+
+// ExtendPeerSession mocks ExtendPeerSession of the AccountManager interface
+func (am *MockAccountManager) ExtendPeerSession(ctx context.Context, peerPubKey, userID string) (time.Time, error) {
+	if am.ExtendPeerSessionFunc != nil {
+		return am.ExtendPeerSessionFunc(ctx, peerPubKey, userID)
+	}
+	return time.Time{}, status.Errorf(codes.Unimplemented, "method ExtendPeerSession is not implemented")
 }
 
 // SyncPeer mocks SyncPeer of the AccountManager interface
-func (am *MockAccountManager) SyncPeer(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error) {
+func (am *MockAccountManager) SyncPeer(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*nmdata.PostureChecks, int64, error) {
 	if am.SyncPeerFunc != nil {
 		return am.SyncPeerFunc(ctx, sync, accountID)
 	}
@@ -934,9 +976,9 @@ func (am *MockAccountManager) GroupValidation(ctx context.Context, accountId str
 }
 
 // SyncPeerMeta mocks SyncPeerMeta of the AccountManager interface
-func (am *MockAccountManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error {
+func (am *MockAccountManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error {
 	if am.SyncPeerMetaFunc != nil {
-		return am.SyncPeerMetaFunc(ctx, peerPubKey, meta)
+		return am.SyncPeerMetaFunc(ctx, peerPubKey, meta, realIP)
 	}
 	return status.Errorf(codes.Unimplemented, "method SyncPeerMeta is not implemented")
 }

management/server/mock_server/management_server_mock.go

@@ -13,7 +13,7 @@ type ManagementServiceServerMock struct {
 	proto.UnimplementedManagementServiceServer
 
 	LoginFunc                      func(context.Context, *proto.EncryptedMessage) (*proto.EncryptedMessage, error)
-	SyncFunc                       func(*proto.EncryptedMessage, proto.ManagementService_SyncServer)
+	SyncFunc                       func(*proto.EncryptedMessage, proto.ManagementService_SyncServer) error
 	GetServerKeyFunc               func(context.Context, *proto.Empty) (*proto.ServerKeyResponse, error)
 	IsHealthyFunc                  func(context.Context, *proto.Empty) (*proto.Empty, error)
 	GetDeviceAuthorizationFlowFunc func(ctx context.Context, req *proto.EncryptedMessage) (*proto.EncryptedMessage, error)
@@ -30,7 +30,7 @@ func (m ManagementServiceServerMock) Login(ctx context.Context, req *proto.Encry
 
 func (m ManagementServiceServerMock) Sync(msg *proto.EncryptedMessage, sync proto.ManagementService_SyncServer) error {
 	if m.SyncFunc != nil {
-		return m.Sync(msg, sync)
+		return m.SyncFunc(msg, sync)
 	}
 	return status.Errorf(codes.Unimplemented, "method Sync not implemented")
 }

management/server/nameserver.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"slices"
 	"strings"
 	"unicode/utf8"
 
@@ -11,6 +12,7 @@ import (
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 	nbdomain "github.com/netbirdio/netbird/shared/management/domain"
@@ -39,7 +41,8 @@ func (am *DefaultAccountManager) CreateNameServerGroup(ctx context.Context, acco
 		SearchDomainsEnabled: searchDomainEnabled,
 	}
 
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
+	change := affectedpeers.Change{DistributionGroupIDs: newNSGroup.Groups}
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		var err error
@@ -47,12 +50,13 @@ func (am *DefaultAccountManager) CreateNameServerGroup(ctx context.Context, acco
 			return err
 		}
 
-		updateAccountPeers, err = anyGroupHasPeersOrResources(ctx, transaction, accountID, newNSGroup.Groups)
-		if err != nil {
+		newNSGroup.PublicID = xid.New().String()
+
+		if err = transaction.SaveNameServerGroup(ctx, newNSGroup); err != nil {
 			return err
 		}
 
-		if err = transaction.SaveNameServerGroup(ctx, newNSGroup); err != nil {
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -64,9 +68,7 @@ func (am *DefaultAccountManager) CreateNameServerGroup(ctx context.Context, acco
 
 	am.StoreEvent(ctx, userID, newNSGroup.ID, accountID, activity.NameserverGroupCreated, newNSGroup.EventMeta())
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return newNSGroup.Copy(), nil
 }
@@ -77,7 +79,8 @@ func (am *DefaultAccountManager) SaveNameServerGroup(ctx context.Context, accoun
 		return status.Errorf(status.InvalidArgument, "nameserver group provided is nil")
 	}
 
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		oldNSGroup, err := transaction.GetNameServerGroupByID(ctx, store.LockingStrengthNone, accountID, nsGroupToSave.ID)
@@ -90,12 +93,14 @@ func (am *DefaultAccountManager) SaveNameServerGroup(ctx context.Context, accoun
 			return err
 		}
 
-		updateAccountPeers, err = areNameServerGroupChangesAffectPeers(ctx, transaction, nsGroupToSave, oldNSGroup)
-		if err != nil {
+		nsGroupToSave.PublicID = oldNSGroup.PublicID
+
+		if err = transaction.SaveNameServerGroup(ctx, nsGroupToSave); err != nil {
 			return err
 		}
 
-		if err = transaction.SaveNameServerGroup(ctx, nsGroupToSave); err != nil {
+		change = affectedpeers.Change{DistributionGroupIDs: slices.Concat(nsGroupToSave.Groups, oldNSGroup.Groups)}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -107,9 +112,7 @@ func (am *DefaultAccountManager) SaveNameServerGroup(ctx context.Context, accoun
 
 	am.StoreEvent(ctx, userID, nsGroupToSave.ID, accountID, activity.NameserverGroupUpdated, nsGroupToSave.EventMeta())
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -117,7 +120,8 @@ func (am *DefaultAccountManager) SaveNameServerGroup(ctx context.Context, accoun
 // DeleteNameServerGroup deletes nameserver group with nsGroupID
 func (am *DefaultAccountManager) DeleteNameServerGroup(ctx context.Context, accountID, nsGroupID, userID string) error {
 	var nsGroup *nbdns.NameServerGroup
-	var updateAccountPeers bool
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		var err error
@@ -126,8 +130,9 @@ func (am *DefaultAccountManager) DeleteNameServerGroup(ctx context.Context, acco
 			return err
 		}
 
-		updateAccountPeers, err = anyGroupHasPeersOrResources(ctx, transaction, accountID, nsGroup.Groups)
-		if err != nil {
+		// Load before delete: the post-delete state no longer references the groups.
+		change = affectedpeers.Change{DistributionGroupIDs: nsGroup.Groups}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err
 		}
 
@@ -143,9 +148,7 @@ func (am *DefaultAccountManager) DeleteNameServerGroup(ctx context.Context, acco
 
 	am.StoreEvent(ctx, userID, nsGroup.ID, accountID, activity.NameserverGroupDeleted, nsGroup.EventMeta())
 
-	if updateAccountPeers {
-		am.UpdateAccountPeers(ctx, accountID)
-	}
+	am.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
@@ -184,24 +187,6 @@ func validateNameServerGroup(ctx context.Context, transaction store.Store, accou
 	return validateGroups(nameserverGroup.Groups, groups)
 }
 
-// areNameServerGroupChangesAffectPeers checks if the changes in the nameserver group affect the peers.
-func areNameServerGroupChangesAffectPeers(ctx context.Context, transaction store.Store, newNSGroup, oldNSGroup *nbdns.NameServerGroup) (bool, error) {
-	if !newNSGroup.Enabled && !oldNSGroup.Enabled {
-		return false, nil
-	}
-
-	hasPeers, err := anyGroupHasPeersOrResources(ctx, transaction, newNSGroup.AccountID, newNSGroup.Groups)
-	if err != nil {
-		return false, err
-	}
-
-	if hasPeers {
-		return true, nil
-	}
-
-	return anyGroupHasPeersOrResources(ctx, transaction, oldNSGroup.AccountID, oldNSGroup.Groups)
-}
-
 func validateDomainInput(primary bool, domains []string, searchDomainsEnabled bool) error {
 	if !primary && len(domains) == 0 {
 		return status.Errorf(status.InvalidArgument, "nameserver group primary status is false and domains are empty,"+

management/server/nameserver_test.go

@@ -6,9 +6,9 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+	"go.uber.org/mock/gomock"
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
@@ -803,7 +803,7 @@ func createNSManager(t *testing.T) (*DefaultAccountManager, error) {
 
 	updateManager := update_channel.NewPeersUpdateManager(metrics)
 	requestBuffer := NewAccountRequestBuffer(ctx, store)
-	networkMapController := controller.NewController(ctx, store, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peers.NewManager(store)), &config.Config{})
+	networkMapController := controller.NewController(ctx, store, metrics, updateManager, requestBuffer, MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), ephemeral_manager.NewEphemeralManager(store, peers.NewManager(store)), &config.Config{}, nil)
 
 	return BuildManager(context.Background(), nil, store, networkMapController, job.NewJobManager(nil, store, peersManager), nil, "", eventStore, nil, false, MockIntegratedValidator{}, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
 }
@@ -896,11 +896,11 @@ func initTestNSAccount(t *testing.T, am *DefaultAccountManager) (*types.Account,
 		return nil, err
 	}
 
-	_, _, _, err = am.AddPeer(context.Background(), "", "", userID, peer1, false)
+	_, _, _, _, err = am.AddPeer(context.Background(), "", "", userID, peer1, false)
 	if err != nil {
 		return nil, err
 	}
-	_, _, _, err = am.AddPeer(context.Background(), "", "", userID, peer2, false)
+	_, _, _, _, err = am.AddPeer(context.Background(), "", "", userID, peer2, false)
 	if err != nil {
 		return nil, err
 	}
@@ -1087,7 +1087,7 @@ func TestNameServerAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})
@@ -1105,7 +1105,7 @@ func TestNameServerAccountPeersUpdate(t *testing.T) {
 
 		select {
 		case <-done:
-		case <-time.After(time.Second):
+		case <-time.After(peerUpdateTimeout):
 			t.Error("timeout waiting for peerShouldReceiveUpdate")
 		}
 	})

management/server/networks/manager.go

@@ -8,6 +8,7 @@ import (
 
 	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	"github.com/netbirdio/netbird/management/server/networks/resources"
 	"github.com/netbirdio/netbird/management/server/networks/routers"
 	"github.com/netbirdio/netbird/management/server/networks/types"
@@ -48,9 +49,16 @@ func (m *managerImpl) GetAllNetworks(ctx context.Context, accountID, userID stri
 func (m *managerImpl) CreateNetwork(ctx context.Context, userID string, network *types.Network) (*types.Network, error) {
 	network.ID = xid.New().String()
 
-	err := m.store.SaveNetwork(ctx, network)
+	err := m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
+		network.PublicID = xid.New().String()
+
+		if err := transaction.SaveNetwork(ctx, network); err != nil {
+			return fmt.Errorf("failed to save network: %w", err)
+		}
+		return nil
+	})
 	if err != nil {
-		return nil, fmt.Errorf("failed to save network: %w", err)
+		return nil, err
 	}
 
 	m.accountManager.StoreEvent(ctx, userID, network.ID, network.AccountID, activity.NetworkCreated, network.EventMeta())
@@ -63,14 +71,25 @@ func (m *managerImpl) GetNetwork(ctx context.Context, accountID, userID, network
 }
 
 func (m *managerImpl) UpdateNetwork(ctx context.Context, userID string, network *types.Network) (*types.Network, error) {
-	_, err := m.store.GetNetworkByID(ctx, store.LockingStrengthUpdate, network.AccountID, network.ID)
+	err := m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
+		existing, err := transaction.GetNetworkByID(ctx, store.LockingStrengthUpdate, network.AccountID, network.ID)
+		if err != nil {
+			return fmt.Errorf("failed to get network: %w", err)
+		}
+		network.PublicID = existing.PublicID
+
+		if err := transaction.SaveNetwork(ctx, network); err != nil {
+			return fmt.Errorf("failed to save network: %w", err)
+		}
+		return nil
+	})
 	if err != nil {
-		return nil, fmt.Errorf("failed to get network: %w", err)
+		return nil, err
 	}
 
 	m.accountManager.StoreEvent(ctx, userID, network.ID, network.AccountID, activity.NetworkUpdated, network.EventMeta())
 
-	return network, m.store.SaveNetwork(ctx, network)
+	return network, nil
 }
 
 func (m *managerImpl) DeleteNetwork(ctx context.Context, accountID, userID, networkID string) error {
@@ -80,30 +99,39 @@ func (m *managerImpl) DeleteNetwork(ctx context.Context, accountID, userID, netw
 	}
 
 	var eventsToStore []func()
+	var snap *affectedpeers.Snapshot
+	change := affectedpeers.Change{Networks: []*types.Network{network}}
 	err = m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		resources, err := transaction.GetNetworkResourcesByNetID(ctx, store.LockingStrengthUpdate, accountID, networkID)
 		if err != nil {
 			return fmt.Errorf("failed to get resources in network: %w", err)
 		}
 
-		for _, resource := range resources {
-			event, err := m.resourcesManager.DeleteResourceInTransaction(ctx, transaction, accountID, userID, networkID, resource.ID)
-			if err != nil {
-				return fmt.Errorf("failed to delete resource: %w", err)
-			}
-			eventsToStore = append(eventsToStore, event...)
-		}
-
-		routers, err := transaction.GetNetworkRoutersByNetID(ctx, store.LockingStrengthUpdate, accountID, networkID)
+		netRouters, err := transaction.GetNetworkRoutersByNetID(ctx, store.LockingStrengthUpdate, accountID, networkID)
 		if err != nil {
 			return fmt.Errorf("failed to get routers in network: %w", err)
 		}
 
-		for _, router := range routers {
-			event, err := m.routersManager.DeleteRouterInTransaction(ctx, transaction, accountID, userID, networkID, router.ID)
+		var lerr error
+		if snap, lerr = affectedpeers.Load(ctx, transaction, accountID, change); lerr != nil {
+			return lerr
+		}
+
+		for _, resource := range resources {
+			deleted, event, err := m.resourcesManager.DeleteResourceInTransaction(ctx, transaction, accountID, userID, networkID, resource.ID)
+			if err != nil {
+				return fmt.Errorf("failed to delete resource: %w", err)
+			}
+			change.Resources = append(change.Resources, deleted)
+			eventsToStore = append(eventsToStore, event...)
+		}
+
+		for _, router := range netRouters {
+			deleted, event, err := m.routersManager.DeleteRouterInTransaction(ctx, transaction, accountID, userID, networkID, router.ID)
 			if err != nil {
 				return fmt.Errorf("failed to delete router: %w", err)
 			}
+			change.Routers = append(change.Routers, deleted)
 			eventsToStore = append(eventsToStore, event)
 		}
 
@@ -131,7 +159,7 @@ func (m *managerImpl) DeleteNetwork(ctx context.Context, accountID, userID, netw
 		event()
 	}
 
-	go m.accountManager.UpdateAccountPeers(ctx, accountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }

management/server/networks/manager_test.go

@@ -32,8 +32,11 @@ func Test_GetAllNetworksReturnsNetworks(t *testing.T) {
 
 	networks, err := manager.GetAllNetworks(ctx, accountID, userID)
 	require.NoError(t, err)
-	require.Len(t, networks, 1)
-	require.Equal(t, "testNetworkId", networks[0].ID)
+	ids := make([]string, 0, len(networks))
+	for _, n := range networks {
+		ids = append(ids, n.ID)
+	}
+	require.ElementsMatch(t, []string{"testNetworkId", "secondNetworkId"}, ids)
 }
 
 func Test_GetNetworkReturnsNetwork(t *testing.T) {
@@ -127,3 +130,68 @@ func Test_UpdateNetworkSuccessfully(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, network.Name, updatedNetwork.Name)
 }
+
+func Test_CreateNetworkSetsPublicId(t *testing.T) {
+	ctx := context.Background()
+	const accountID = "testAccountId"
+	const userID = "testAdminId"
+
+	s, cleanUp, err := store.NewTestStoreFromSQL(ctx, "../testdata/networks.sql", t.TempDir())
+	require.NoError(t, err)
+	t.Cleanup(cleanUp)
+
+	am := mock_server.MockAccountManager{}
+	groupsManager := groups.NewManagerMock()
+	routerManager := routers.NewManagerMock()
+	resourcesManager := resources.NewManager(s, groupsManager, &am, nil)
+	manager := NewManager(s, resourcesManager, routerManager, &am)
+
+	created, err := manager.CreateNetwork(ctx, userID, &types.Network{
+		AccountID: accountID,
+		Name:      "seq-allocation-test",
+	})
+	require.NoError(t, err)
+	require.NotEqual(t, "", created.PublicID, "CreateNetwork must allocate a non-zero AccountSeqID")
+}
+
+// Test_UpdateNetworkPreservesSeqID verifies UpdateNetwork does not reset
+// AccountSeqID even when the caller passes a zero value (the shape REST
+// handlers produce because the field is `json:"-"`).
+func Test_UpdateNetworkPreservesPublicId(t *testing.T) {
+	ctx := context.Background()
+	const accountID = "testAccountId"
+	const userID = "testAdminId"
+
+	s, cleanUp, err := store.NewTestStoreFromSQL(ctx, "../testdata/networks.sql", t.TempDir())
+	require.NoError(t, err)
+	t.Cleanup(cleanUp)
+
+	am := mock_server.MockAccountManager{}
+	groupsManager := groups.NewManagerMock()
+	routerManager := routers.NewManagerMock()
+	resourcesManager := resources.NewManager(s, groupsManager, &am, nil)
+	manager := NewManager(s, resourcesManager, routerManager, &am)
+
+	created, err := manager.CreateNetwork(ctx, userID, &types.Network{
+		AccountID: accountID,
+		Name:      "seq-preserve-original",
+	})
+	require.NoError(t, err)
+	originalPublicId := created.PublicID
+	require.NotZero(t, originalPublicId)
+
+	update := &types.Network{
+		AccountID: accountID,
+		ID:        created.ID,
+		Name:      "seq-preserve-renamed",
+	}
+	require.Equal(t, "", update.PublicID, "incoming struct must mirror an HTTP handler shape")
+
+	_, err = manager.UpdateNetwork(ctx, userID, update)
+	require.NoError(t, err)
+
+	got, err := manager.GetNetwork(ctx, accountID, userID, created.ID)
+	require.NoError(t, err)
+	require.Equal(t, originalPublicId, got.PublicID, "PublicID must survive UpdateNetwork")
+	require.Equal(t, "seq-preserve-renamed", got.Name)
+}

management/server/networks/resources/manager.go

@@ -5,11 +5,13 @@ import (
 	"errors"
 	"fmt"
 
+	"github.com/rs/xid"
 	log "github.com/sirupsen/logrus"
 
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	"github.com/netbirdio/netbird/management/server/groups"
 	"github.com/netbirdio/netbird/management/server/networks/resources/types"
 	"github.com/netbirdio/netbird/management/server/store"
@@ -26,7 +28,7 @@ type Manager interface {
 	GetResource(ctx context.Context, accountID, userID, networkID, resourceID string) (*types.NetworkResource, error)
 	UpdateResource(ctx context.Context, userID string, resource *types.NetworkResource) (*types.NetworkResource, error)
 	DeleteResource(ctx context.Context, accountID, userID, networkID, resourceID string) error
-	DeleteResourceInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, resourceID string) ([]func(), error)
+	DeleteResourceInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, resourceID string) (*types.NetworkResource, []func(), error)
 }
 
 type managerImpl struct {
@@ -77,45 +79,12 @@ func (m *managerImpl) CreateResource(ctx context.Context, userID string, resourc
 	}
 
 	var eventsToStore []func()
+	var snap *affectedpeers.Snapshot
+	change := affectedpeers.Change{Resources: []*types.NetworkResource{resource}}
 	err = m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		_, err = transaction.GetNetworkResourceByName(ctx, store.LockingStrengthNone, resource.AccountID, resource.Name)
-		if err == nil {
-			return status.Errorf(status.InvalidArgument, "resource with name %s already exists", resource.Name)
-		}
-
-		network, err := transaction.GetNetworkByID(ctx, store.LockingStrengthUpdate, resource.AccountID, resource.NetworkID)
-		if err != nil {
-			return fmt.Errorf("failed to get network: %w", err)
-		}
-
-		err = transaction.SaveNetworkResource(ctx, resource)
-		if err != nil {
-			return fmt.Errorf("failed to save network resource: %w", err)
-		}
-
-		event := func() {
-			m.accountManager.StoreEvent(ctx, userID, resource.ID, resource.AccountID, activity.NetworkResourceCreated, resource.EventMeta(network))
-		}
-		eventsToStore = append(eventsToStore, event)
-
-		res := nbtypes.Resource{
-			ID:   resource.ID,
-			Type: nbtypes.ResourceType(resource.Type.String()),
-		}
-		for _, groupID := range resource.GroupIDs {
-			event, err := m.groupsManager.AddResourceToGroupInTransaction(ctx, transaction, resource.AccountID, userID, groupID, &res)
-			if err != nil {
-				return fmt.Errorf("failed to add resource to group: %w", err)
-			}
-			eventsToStore = append(eventsToStore, event)
-		}
-
-		err = transaction.IncrementNetworkSerial(ctx, resource.AccountID)
-		if err != nil {
-			return fmt.Errorf("failed to increment network serial: %w", err)
-		}
-
-		return nil
+		var txErr error
+		eventsToStore, snap, txErr = m.createResourceInTransaction(ctx, transaction, userID, resource, change)
+		return txErr
 	})
 	if err != nil {
 		return nil, fmt.Errorf("failed to create network resource: %w", err)
@@ -125,11 +94,57 @@ func (m *managerImpl) CreateResource(ctx context.Context, userID string, resourc
 		event()
 	}
 
-	go m.accountManager.UpdateAccountPeers(ctx, resource.AccountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, resource.AccountID, snap, change)
 
 	return resource, nil
 }
 
+func (m *managerImpl) createResourceInTransaction(ctx context.Context, transaction store.Store, userID string, resource *types.NetworkResource, change affectedpeers.Change) ([]func(), *affectedpeers.Snapshot, error) {
+	_, err := transaction.GetNetworkResourceByName(ctx, store.LockingStrengthNone, resource.AccountID, resource.Name)
+	if err == nil {
+		return nil, nil, status.Errorf(status.InvalidArgument, "resource with name %s already exists", resource.Name)
+	}
+
+	network, err := transaction.GetNetworkByID(ctx, store.LockingStrengthUpdate, resource.AccountID, resource.NetworkID)
+	if err != nil {
+		return nil, nil, fmt.Errorf("failed to get network: %w", err)
+	}
+
+	resource.PublicID = xid.New().String()
+
+	if err = transaction.SaveNetworkResource(ctx, resource); err != nil {
+		return nil, nil, fmt.Errorf("failed to save network resource: %w", err)
+	}
+
+	var eventsToStore []func()
+	eventsToStore = append(eventsToStore, func() {
+		m.accountManager.StoreEvent(ctx, userID, resource.ID, resource.AccountID, activity.NetworkResourceCreated, resource.EventMeta(network))
+	})
+
+	res := nbtypes.Resource{
+		ID:   resource.ID,
+		Type: nbtypes.ResourceType(resource.Type.String()),
+	}
+	for _, groupID := range resource.GroupIDs {
+		event, err := m.groupsManager.AddResourceToGroupInTransaction(ctx, transaction, resource.AccountID, userID, groupID, &res)
+		if err != nil {
+			return nil, nil, fmt.Errorf("failed to add resource to group: %w", err)
+		}
+		eventsToStore = append(eventsToStore, event)
+	}
+
+	if err = transaction.IncrementNetworkSerial(ctx, resource.AccountID); err != nil {
+		return nil, nil, fmt.Errorf("failed to increment network serial: %w", err)
+	}
+
+	snap, err := affectedpeers.Load(ctx, transaction, resource.AccountID, change)
+	if err != nil {
+		return nil, nil, err
+	}
+
+	return eventsToStore, snap, nil
+}
+
 func (m *managerImpl) GetResource(ctx context.Context, accountID, userID, networkID, resourceID string) (*types.NetworkResource, error) {
 	resource, err := m.store.GetNetworkResourceByID(ctx, store.LockingStrengthNone, accountID, resourceID)
 	if err != nil {
@@ -154,6 +169,8 @@ func (m *managerImpl) UpdateResource(ctx context.Context, userID string, resourc
 	resource.Prefix = prefix
 
 	var eventsToStore []func()
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 	err = m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		network, err := transaction.GetNetworkByID(ctx, store.LockingStrengthUpdate, resource.AccountID, resource.NetworkID)
 		if err != nil {
@@ -178,6 +195,15 @@ func (m *managerImpl) UpdateResource(ctx context.Context, userID string, resourc
 		if err != nil {
 			return fmt.Errorf("failed to get network resource: %w", err)
 		}
+		resource.PublicID = oldResource.PublicID
+
+		oldGroups, err := m.groupsManager.GetResourceGroupsInTransaction(ctx, transaction, store.LockingStrengthNone, resource.AccountID, resource.ID)
+		if err != nil {
+			return fmt.Errorf("failed to get old resource groups: %w", err)
+		}
+		for _, g := range oldGroups {
+			oldResource.GroupIDs = append(oldResource.GroupIDs, g.ID)
+		}
 
 		err = transaction.SaveNetworkResource(ctx, resource)
 		if err != nil {
@@ -194,6 +220,11 @@ func (m *managerImpl) UpdateResource(ctx context.Context, userID string, resourc
 			m.accountManager.StoreEvent(ctx, userID, resource.ID, resource.AccountID, activity.NetworkResourceUpdated, resource.EventMeta(network))
 		})
 
+		change = affectedpeers.Change{Resources: []*types.NetworkResource{oldResource, resource}}
+		if snap, err = affectedpeers.Load(ctx, transaction, resource.AccountID, change); err != nil {
+			return err
+		}
+
 		err = transaction.IncrementNetworkSerial(ctx, resource.AccountID)
 		if err != nil {
 			return fmt.Errorf("failed to increment network serial: %w", err)
@@ -217,7 +248,7 @@ func (m *managerImpl) UpdateResource(ctx context.Context, userID string, resourc
 		}
 	}()
 
-	go m.accountManager.UpdateAccountPeers(ctx, resource.AccountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, resource.AccountID, snap, change)
 
 	return resource, nil
 }
@@ -270,8 +301,26 @@ func (m *managerImpl) DeleteResource(ctx context.Context, accountID, userID, net
 	}
 
 	var events []func()
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 	err = m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		events, err = m.DeleteResourceInTransaction(ctx, transaction, accountID, userID, networkID, resourceID)
+		existing, err := transaction.GetNetworkResourceByID(ctx, store.LockingStrengthUpdate, accountID, resourceID)
+		if err != nil {
+			return fmt.Errorf("failed to get network resource: %w", err)
+		}
+		oldGroups, err := m.groupsManager.GetResourceGroupsInTransaction(ctx, transaction, store.LockingStrengthNone, accountID, resourceID)
+		if err != nil {
+			return fmt.Errorf("failed to get resource groups: %w", err)
+		}
+		for _, g := range oldGroups {
+			existing.GroupIDs = append(existing.GroupIDs, g.ID)
+		}
+		change = affectedpeers.Change{Resources: []*types.NetworkResource{existing}}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
+			return err
+		}
+
+		_, events, err = m.DeleteResourceInTransaction(ctx, transaction, accountID, userID, networkID, resourceID)
 		if err != nil {
 			return fmt.Errorf("failed to delete resource: %w", err)
 		}
@@ -291,51 +340,53 @@ func (m *managerImpl) DeleteResource(ctx context.Context, accountID, userID, net
 		event()
 	}
 
-	go m.accountManager.UpdateAccountPeers(ctx, accountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
 
-func (m *managerImpl) DeleteResourceInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, resourceID string) ([]func(), error) {
+func (m *managerImpl) DeleteResourceInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, resourceID string) (*types.NetworkResource, []func(), error) {
 	resource, err := transaction.GetNetworkResourceByID(ctx, store.LockingStrengthUpdate, accountID, resourceID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to get network resource: %w", err)
+		return nil, nil, fmt.Errorf("failed to get network resource: %w", err)
 	}
 
 	network, err := transaction.GetNetworkByID(ctx, store.LockingStrengthUpdate, accountID, networkID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to get network: %w", err)
+		return nil, nil, fmt.Errorf("failed to get network: %w", err)
 	}
 
 	if resource.NetworkID != networkID {
-		return nil, errors.New("resource not part of network")
+		return nil, nil, errors.New("resource not part of network")
 	}
 
 	groups, err := m.groupsManager.GetResourceGroupsInTransaction(ctx, transaction, store.LockingStrengthUpdate, accountID, resourceID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to get resource groups: %w", err)
+		return nil, nil, fmt.Errorf("failed to get resource groups: %w", err)
 	}
 
 	var eventsToStore []func()
 
 	for _, group := range groups {
+		resource.GroupIDs = append(resource.GroupIDs, group.ID)
+
 		event, err := m.groupsManager.RemoveResourceFromGroupInTransaction(ctx, transaction, accountID, userID, group.ID, resourceID)
 		if err != nil {
-			return nil, fmt.Errorf("failed to remove resource from group: %w", err)
+			return nil, nil, fmt.Errorf("failed to remove resource from group: %w", err)
 		}
 		eventsToStore = append(eventsToStore, event)
 	}
 
 	err = transaction.DeleteNetworkResource(ctx, accountID, resourceID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to delete network resource: %w", err)
+		return nil, nil, fmt.Errorf("failed to delete network resource: %w", err)
 	}
 
 	eventsToStore = append(eventsToStore, func() {
 		m.accountManager.StoreEvent(ctx, userID, resourceID, accountID, activity.NetworkResourceDeleted, resource.EventMeta(network))
 	})
 
-	return eventsToStore, nil
+	return resource, eventsToStore, nil
 }
 
 func NewManagerMock() Manager {
@@ -370,6 +421,6 @@ func (m *mockManager) DeleteResource(ctx context.Context, accountID, userID, net
 	return nil
 }
 
-func (m *mockManager) DeleteResourceInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, resourceID string) ([]func(), error) {
-	return []func(){}, nil
+func (m *mockManager) DeleteResourceInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, resourceID string) (*types.NetworkResource, []func(), error) {
+	return nil, []func(){}, nil
 }

management/server/networks/resources/manager_test.go

@@ -4,7 +4,7 @@ import (
 	"context"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/require"
 
 	reverseproxy "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"

management/server/networks/resources/types/resource.go

@@ -32,6 +32,7 @@ type NetworkResource struct {
 	ID          string `gorm:"primaryKey"`
 	NetworkID   string `gorm:"index"`
 	AccountID   string `gorm:"index"`
+	PublicID    string `json:"-"`
 	Name        string
 	Description string
 	Type        NetworkResourceType
@@ -96,6 +97,7 @@ func (n *NetworkResource) Copy() *NetworkResource {
 		ID:          n.ID,
 		AccountID:   n.AccountID,
 		NetworkID:   n.NetworkID,
+		PublicID:    n.PublicID,
 		Name:        n.Name,
 		Description: n.Description,
 		Type:        n.Type,

management/server/networks/routers/manager.go

@@ -9,6 +9,7 @@ import (
 
 	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	"github.com/netbirdio/netbird/management/server/networks/routers/types"
 	networkTypes "github.com/netbirdio/netbird/management/server/networks/types"
 	"github.com/netbirdio/netbird/management/server/store"
@@ -22,7 +23,7 @@ type Manager interface {
 	GetRouter(ctx context.Context, accountID, userID, networkID, routerID string) (*types.NetworkRouter, error)
 	UpdateRouter(ctx context.Context, userID string, router *types.NetworkRouter) (*types.NetworkRouter, error)
 	DeleteRouter(ctx context.Context, accountID, userID, networkID, routerID string) error
-	DeleteRouterInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, routerID string) (func(), error)
+	DeleteRouterInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, routerID string) (*types.NetworkRouter, func(), error)
 }
 
 type managerImpl struct {
@@ -60,8 +61,10 @@ func (m *managerImpl) GetAllRoutersInAccount(ctx context.Context, accountID, use
 
 func (m *managerImpl) CreateRouter(ctx context.Context, userID string, router *types.NetworkRouter) (*types.NetworkRouter, error) {
 	var network *networkTypes.Network
-	err := m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		var err error
+	var snap *affectedpeers.Snapshot
+	change := affectedpeers.Change{Routers: []*types.NetworkRouter{router}}
+	var err error
+	err = m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		network, err = transaction.GetNetworkByID(ctx, store.LockingStrengthNone, router.AccountID, router.NetworkID)
 		if err != nil {
 			return fmt.Errorf("failed to get network: %w", err)
@@ -73,7 +76,9 @@ func (m *managerImpl) CreateRouter(ctx context.Context, userID string, router *t
 
 		router.ID = xid.New().String()
 
-		err = transaction.SaveNetworkRouter(ctx, router)
+		router.PublicID = xid.New().String()
+
+		err = transaction.CreateNetworkRouter(ctx, router)
 		if err != nil {
 			return fmt.Errorf("failed to create network router: %w", err)
 		}
@@ -83,6 +88,10 @@ func (m *managerImpl) CreateRouter(ctx context.Context, userID string, router *t
 			return fmt.Errorf("failed to increment network serial: %w", err)
 		}
 
+		if snap, err = affectedpeers.Load(ctx, transaction, router.AccountID, change); err != nil {
+			return err
+		}
+
 		return nil
 	})
 	if err != nil {
@@ -91,7 +100,7 @@ func (m *managerImpl) CreateRouter(ctx context.Context, userID string, router *t
 
 	m.accountManager.StoreEvent(ctx, userID, router.ID, router.AccountID, activity.NetworkRouterCreated, router.EventMeta(network))
 
-	go m.accountManager.UpdateAccountPeers(ctx, router.AccountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, router.AccountID, snap, change)
 
 	return router, nil
 }
@@ -111,28 +120,12 @@ func (m *managerImpl) GetRouter(ctx context.Context, accountID, userID, networkI
 
 func (m *managerImpl) UpdateRouter(ctx context.Context, userID string, router *types.NetworkRouter) (*types.NetworkRouter, error) {
 	var network *networkTypes.Network
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 	err := m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		var err error
-		network, err = transaction.GetNetworkByID(ctx, store.LockingStrengthNone, router.AccountID, router.NetworkID)
-		if err != nil {
-			return fmt.Errorf("failed to get network: %w", err)
-		}
-
-		if network.ID != router.NetworkID {
-			return status.NewRouterNotPartOfNetworkError(router.ID, router.NetworkID)
-		}
-
-		err = transaction.SaveNetworkRouter(ctx, router)
-		if err != nil {
-			return fmt.Errorf("failed to update network router: %w", err)
-		}
-
-		err = transaction.IncrementNetworkSerial(ctx, router.AccountID)
-		if err != nil {
-			return fmt.Errorf("failed to increment network serial: %w", err)
-		}
-
-		return nil
+		var txErr error
+		network, snap, change, txErr = m.updateRouterInTransaction(ctx, transaction, router)
+		return txErr
 	})
 	if err != nil {
 		return nil, err
@@ -140,16 +133,67 @@ func (m *managerImpl) UpdateRouter(ctx context.Context, userID string, router *t
 
 	m.accountManager.StoreEvent(ctx, userID, router.ID, router.AccountID, activity.NetworkRouterUpdated, router.EventMeta(network))
 
-	go m.accountManager.UpdateAccountPeers(ctx, router.AccountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, router.AccountID, snap, change)
 
 	return router, nil
 }
 
+func (m *managerImpl) updateRouterInTransaction(ctx context.Context, transaction store.Store, router *types.NetworkRouter) (*networkTypes.Network, *affectedpeers.Snapshot, affectedpeers.Change, error) {
+	network, err := transaction.GetNetworkByID(ctx, store.LockingStrengthNone, router.AccountID, router.NetworkID)
+	if err != nil {
+		return nil, nil, affectedpeers.Change{}, fmt.Errorf("failed to get network: %w", err)
+	}
+
+	existing, err := transaction.GetNetworkRouterByID(ctx, store.LockingStrengthUpdate, router.AccountID, router.ID)
+	if err != nil {
+		return nil, nil, affectedpeers.Change{}, fmt.Errorf("failed to get network router: %w", err)
+	}
+
+	if existing.AccountID != router.AccountID {
+		return nil, nil, affectedpeers.Change{}, status.NewNetworkRouterNotFoundError(router.ID)
+	}
+
+	if existing.NetworkID != router.NetworkID {
+		return nil, nil, affectedpeers.Change{}, status.NewRouterNotPartOfNetworkError(router.ID, router.NetworkID)
+	}
+
+	// Preserve PublicID from the existing router so the upstream
+	// UpdateNetworkRouter (which does Updates(router) with Select("*"))
+	// doesn't clobber it with the request's zero value.
+	router.PublicID = existing.PublicID
+
+	if err = transaction.UpdateNetworkRouter(ctx, router); err != nil {
+		return nil, nil, affectedpeers.Change{}, fmt.Errorf("failed to update network router: %w", err)
+	}
+
+	if err = transaction.IncrementNetworkSerial(ctx, router.AccountID); err != nil {
+		return nil, nil, affectedpeers.Change{}, fmt.Errorf("failed to increment network serial: %w", err)
+	}
+
+	change := affectedpeers.Change{Routers: []*types.NetworkRouter{existing, router}}
+	snap, err := affectedpeers.Load(ctx, transaction, router.AccountID, change)
+	if err != nil {
+		return nil, nil, affectedpeers.Change{}, err
+	}
+
+	return network, snap, change, nil
+}
+
 func (m *managerImpl) DeleteRouter(ctx context.Context, accountID, userID, networkID, routerID string) error {
 	var event func()
+	var snap *affectedpeers.Snapshot
+	var change affectedpeers.Change
 	err := m.store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
-		var err error
-		event, err = m.DeleteRouterInTransaction(ctx, transaction, accountID, userID, networkID, routerID)
+		existing, err := transaction.GetNetworkRouterByID(ctx, store.LockingStrengthUpdate, accountID, routerID)
+		if err != nil {
+			return fmt.Errorf("failed to get network router: %w", err)
+		}
+		change = affectedpeers.Change{Routers: []*types.NetworkRouter{existing}}
+		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
+			return err
+		}
+
+		_, event, err = m.DeleteRouterInTransaction(ctx, transaction, accountID, userID, networkID, routerID)
 		if err != nil {
 			return fmt.Errorf("failed to delete network router: %w", err)
 		}
@@ -167,36 +211,36 @@ func (m *managerImpl) DeleteRouter(ctx context.Context, accountID, userID, netwo
 
 	event()
 
-	go m.accountManager.UpdateAccountPeers(ctx, accountID)
+	m.accountManager.ExpandAndUpdateAffected(ctx, accountID, snap, change)
 
 	return nil
 }
 
-func (m *managerImpl) DeleteRouterInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, routerID string) (func(), error) {
+func (m *managerImpl) DeleteRouterInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, routerID string) (*types.NetworkRouter, func(), error) {
 	network, err := transaction.GetNetworkByID(ctx, store.LockingStrengthNone, accountID, networkID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to get network: %w", err)
+		return nil, nil, fmt.Errorf("failed to get network: %w", err)
 	}
 
 	router, err := transaction.GetNetworkRouterByID(ctx, store.LockingStrengthUpdate, accountID, routerID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to get network router: %w", err)
+		return nil, nil, fmt.Errorf("failed to get network router: %w", err)
 	}
 
 	if router.NetworkID != networkID {
-		return nil, status.NewRouterNotPartOfNetworkError(routerID, networkID)
+		return nil, nil, status.NewRouterNotPartOfNetworkError(routerID, networkID)
 	}
 
 	err = transaction.DeleteNetworkRouter(ctx, accountID, routerID)
 	if err != nil {
-		return nil, fmt.Errorf("failed to delete network router: %w", err)
+		return nil, nil, fmt.Errorf("failed to delete network router: %w", err)
 	}
 
 	event := func() {
 		m.accountManager.StoreEvent(ctx, userID, routerID, accountID, activity.NetworkRouterDeleted, router.EventMeta(network))
 	}
 
-	return event, nil
+	return router, event, nil
 }
 
 func NewManagerMock() Manager {
@@ -227,6 +271,9 @@ func (m *mockManager) DeleteRouter(ctx context.Context, accountID, userID, netwo
 	return nil
 }
 
-func (m *mockManager) DeleteRouterInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, routerID string) (func(), error) {
-	return func() {}, nil
+func (m *mockManager) DeleteRouterInTransaction(ctx context.Context, transaction store.Store, accountID, userID, networkID, routerID string) (*types.NetworkRouter, func(), error) {
+	return nil, func() {
+		// no-op mock: returns zero values so tests that don't exercise router deletion
+		// can satisfy the Manager interface without a real store.
+	}, nil
 }