Merge branch 'main' into embedded-vnc

management/internals/controllers/network_map/controller/controller.go

@@ -176,6 +176,7 @@ func (c *Controller) sendUpdateAccountPeers(ctx context.Context, accountID strin
 	semaphore := make(chan struct{}, 10)
 
 	c.injectAllProxyPolicies(ctx, account)
+	account.PrecomputePostureValidation(ctx)
 	dnsCache := &cache.DNSConfigCache{}
 	dnsDomain := c.GetDNSDomain(account.Settings)
 	peersCustomZone := account.GetPeersCustomZone(ctx, dnsDomain)
@@ -357,6 +358,7 @@ func (c *Controller) sendUpdateForAffectedPeers(ctx context.Context, accountID s
 	// network map that omitted the synth DNS zone, and the agent kept
 	// resolving against the stale or absent record.
 	c.injectAllProxyPolicies(ctx, account)
+	account.PrecomputePostureValidation(ctx)
 	dnsCache := &cache.DNSConfigCache{}
 	dnsDomain := c.GetDNSDomain(account.Settings)
 	peersCustomZone := account.GetPeersCustomZone(ctx, dnsDomain)
@@ -649,6 +651,11 @@ func (c *Controller) GetValidatedPeerWithComponents(ctx context.Context, isRequi
 		return nil, nil, nil, nil, 0, err
 	}
 
+	// it's possible that the peer gets deleted between the call to "sendInitialSync()" and here, bail out in this case
+	if _, ok := account.Peers[peer.ID]; !ok {
+		return nil, nil, nil, nil, 0, fmt.Errorf("peer '%s' no longer exists", peer.ID)
+	}
+
 	c.injectAllProxyPolicies(ctx, account)
 
 	approvedPeersMap, err := c.integratedPeerValidator.GetValidatedPeers(ctx, account.Id, maps.Values(account.Groups), maps.Values(account.Peers), account.Settings.Extra)
@@ -1022,7 +1029,7 @@ func (c *Controller) OnPeersDeleted(ctx context.Context, accountID string, peerI
 					FirewallRules:        []*proto.FirewallRule{},
 					FirewallRulesIsEmpty: true,
 					DNSConfig: &proto.DNSConfig{
-						ForwarderPort: dnsFwdPort,
+						ForwarderPort: dnsFwdPort, //nolint:staticcheck
 					},
 				},
 			},

management/internals/controllers/network_map/controller/controller_test.go

@@ -1,10 +1,15 @@
 package controller
 
 import (
+	"context"
 	"testing"
 
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map"
+	"github.com/netbirdio/netbird/management/server/account"
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/stretchr/testify/assert"
+	"go.uber.org/mock/gomock"
 )
 
 func TestComputeForwarderPort(t *testing.T) {
@@ -107,3 +112,22 @@ func TestComputeForwarderPort(t *testing.T) {
 		t.Errorf("Expected %d for peers with unknown version, got %d", network_map.OldForwarderPort, result)
 	}
 }
+
+func TestGetValidatedPeerWithComponents_DeletedPeer(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mockrequestBuffer := account.NewMockRequestBuffer(ctrl)
+
+	c := Controller{
+		requestBuffer: mockrequestBuffer,
+	}
+
+	mockrequestBuffer.EXPECT().GetAccountWithBackpressure(gomock.Any(), gomock.Any()).Return(&types.Account{}, nil)
+	peer, components, netmap, posturechecks, dnsforwardPort, err := c.GetValidatedPeerWithComponents(context.TODO(), false, "test-account-id", &nbpeer.Peer{ID: "test-peer-id"})
+
+	assert.Nil(t, peer)
+	assert.Nil(t, components)
+	assert.Nil(t, netmap)
+	assert.Nil(t, posturechecks)
+	assert.Equal(t, int64(0), dnsforwardPort)
+	assert.NotNil(t, err)
+}

management/internals/controllers/network_map/controller/repository.go

@@ -3,14 +3,16 @@ package controller
 import (
 	"context"
 
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/netbirdio/netbird/management/internals/modules/zones"
-	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 )
 
+//go:generate go tool mockgen -source=./repository.go -package=controller -destination=repository_mock.go
+
 type Repository interface {
 	GetAccountNetwork(ctx context.Context, accountID string) (*types.Network, error)
 	GetAccountPeers(ctx context.Context, accountID string) ([]*peer.Peer, error)

management/internals/controllers/network_map/controller/repository_mock.go

@@ -0,0 +1,150 @@
+// Code generated by MockGen. DO NOT EDIT.
+// Source: ./repository.go
+//
+// Generated by this command:
+//
+//	mockgen -source=./repository.go -package=controller -destination=repository_mock.go
+//
+
+// Package controller is a generated GoMock package.
+package controller
+
+import (
+	context "context"
+	reflect "reflect"
+
+	service "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
+	zones "github.com/netbirdio/netbird/management/internals/modules/zones"
+	peer "github.com/netbirdio/netbird/management/server/peer"
+	types "github.com/netbirdio/netbird/management/server/types"
+	gomock "go.uber.org/mock/gomock"
+)
+
+// MockRepository is a mock of Repository interface.
+type MockRepository struct {
+	ctrl     *gomock.Controller
+	recorder *MockRepositoryMockRecorder
+	isgomock struct{}
+}
+
+// MockRepositoryMockRecorder is the mock recorder for MockRepository.
+type MockRepositoryMockRecorder struct {
+	mock *MockRepository
+}
+
+// NewMockRepository creates a new mock instance.
+func NewMockRepository(ctrl *gomock.Controller) *MockRepository {
+	mock := &MockRepository{ctrl: ctrl}
+	mock.recorder = &MockRepositoryMockRecorder{mock}
+	return mock
+}
+
+// EXPECT returns an object that allows the caller to indicate expected use.
+func (m *MockRepository) EXPECT() *MockRepositoryMockRecorder {
+	return m.recorder
+}
+
+// GetAccountByPeerID mocks base method.
+func (m *MockRepository) GetAccountByPeerID(ctx context.Context, peerID string) (*types.Account, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountByPeerID", ctx, peerID)
+	ret0, _ := ret[0].(*types.Account)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountByPeerID indicates an expected call of GetAccountByPeerID.
+func (mr *MockRepositoryMockRecorder) GetAccountByPeerID(ctx, peerID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountByPeerID", reflect.TypeOf((*MockRepository)(nil).GetAccountByPeerID), ctx, peerID)
+}
+
+// GetAccountNetwork mocks base method.
+func (m *MockRepository) GetAccountNetwork(ctx context.Context, accountID string) (*types.Network, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountNetwork", ctx, accountID)
+	ret0, _ := ret[0].(*types.Network)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountNetwork indicates an expected call of GetAccountNetwork.
+func (mr *MockRepositoryMockRecorder) GetAccountNetwork(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountNetwork", reflect.TypeOf((*MockRepository)(nil).GetAccountNetwork), ctx, accountID)
+}
+
+// GetAccountPeers mocks base method.
+func (m *MockRepository) GetAccountPeers(ctx context.Context, accountID string) ([]*peer.Peer, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountPeers", ctx, accountID)
+	ret0, _ := ret[0].([]*peer.Peer)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountPeers indicates an expected call of GetAccountPeers.
+func (mr *MockRepositoryMockRecorder) GetAccountPeers(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountPeers", reflect.TypeOf((*MockRepository)(nil).GetAccountPeers), ctx, accountID)
+}
+
+// GetAccountZones mocks base method.
+func (m *MockRepository) GetAccountZones(ctx context.Context, accountID string) ([]*zones.Zone, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountZones", ctx, accountID)
+	ret0, _ := ret[0].([]*zones.Zone)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountZones indicates an expected call of GetAccountZones.
+func (mr *MockRepositoryMockRecorder) GetAccountZones(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountZones", reflect.TypeOf((*MockRepository)(nil).GetAccountZones), ctx, accountID)
+}
+
+// GetPeerByID mocks base method.
+func (m *MockRepository) GetPeerByID(ctx context.Context, accountID, peerID string) (*peer.Peer, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetPeerByID", ctx, accountID, peerID)
+	ret0, _ := ret[0].(*peer.Peer)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetPeerByID indicates an expected call of GetPeerByID.
+func (mr *MockRepositoryMockRecorder) GetPeerByID(ctx, accountID, peerID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeerByID", reflect.TypeOf((*MockRepository)(nil).GetPeerByID), ctx, accountID, peerID)
+}
+
+// GetPeersByIDs mocks base method.
+func (m *MockRepository) GetPeersByIDs(ctx context.Context, accountID string, peerIDs []string) (map[string]*peer.Peer, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetPeersByIDs", ctx, accountID, peerIDs)
+	ret0, _ := ret[0].(map[string]*peer.Peer)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetPeersByIDs indicates an expected call of GetPeersByIDs.
+func (mr *MockRepositoryMockRecorder) GetPeersByIDs(ctx, accountID, peerIDs any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeersByIDs", reflect.TypeOf((*MockRepository)(nil).GetPeersByIDs), ctx, accountID, peerIDs)
+}
+
+// SynthesizeAgentNetworkServices mocks base method.
+func (m *MockRepository) SynthesizeAgentNetworkServices(ctx context.Context, accountID string) ([]*service.Service, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "SynthesizeAgentNetworkServices", ctx, accountID)
+	ret0, _ := ret[0].([]*service.Service)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// SynthesizeAgentNetworkServices indicates an expected call of SynthesizeAgentNetworkServices.
+func (mr *MockRepositoryMockRecorder) SynthesizeAgentNetworkServices(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SynthesizeAgentNetworkServices", reflect.TypeOf((*MockRepository)(nil).SynthesizeAgentNetworkServices), ctx, accountID)
+}

management/internals/controllers/network_map/interface.go

@@ -1,6 +1,6 @@
 package network_map
 
-//go:generate go run go.uber.org/mock/mockgen -package network_map -destination=interface_mock.go -source=./interface.go -build_flags=-mod=mod
+//go:generate go tool mockgen -package network_map -destination=interface_mock.go -source=./interface.go -build_flags=-mod=mod
 
 import (
 	"context"

management/internals/modules/agentnetwork/catalog/catalog.go

@@ -113,8 +113,61 @@ type Provider struct {
 	// upstream provider + credentials on Portkey's hosted side).
 	ExtraHeaders []ExtraHeader
 	Models       []Model
+	// Discovery, when non-nil, describes how to ask this vendor which
+	// models the operator's own credential can actually reach, so the
+	// provider form can offer a live list instead of only the hand-curated
+	// Models above. Nil for entries with no listing endpoint (gateways
+	// vary too much) — those keep free-text entry.
+	Discovery *Discovery
 }
 
+// ListingShape names the response envelope a vendor returns its model
+// listing in. Every vendor invented its own, and none of them can be
+// guessed from the request, so the catalog states it.
+type ListingShape string
+
+const (
+	// ShapeOpenAIData is {"data":[{"id":…}]} — OpenAI, and Anthropic, which
+	// adopted the same envelope.
+	ShapeOpenAIData ListingShape = "openai_data"
+	// ShapeBedrockInferenceProfiles is
+	// {"inferenceProfileSummaries":[{"inferenceProfileId":…}]}. The ids carry
+	// the region prefix that makes them invocable, which is exactly what an
+	// operator cannot reconstruct by hand.
+	ShapeBedrockInferenceProfiles ListingShape = "bedrock_inference_profiles"
+	// ShapeVertexPublisherModels is {"publisherModels":[{"name":…}]}, where
+	// name is a resource path and the invocable id is its last segment joined
+	// to a separate versionId field.
+	ShapeVertexPublisherModels ListingShape = "vertex_publisher_models"
+)
+
+// Discovery describes one vendor's model-listing endpoint.
+//
+// Host is deliberately separate from the provider record's upstream URL:
+// Bedrock serves listings from the control plane (bedrock.<region>) while
+// inference must go to the runtime host (bedrock-runtime.<region>), so the
+// two cannot be the same value. Empty Host means "use the record's own
+// upstream", which is right for every vendor that serves both from one host.
+//
+// The regionPlaceholder in Host is substituted from the provider record's
+// region. Deriving the discovery host from the catalog rather than accepting
+// one from the caller is also what keeps this from being an open proxy: the
+// only hosts management will dial are the ones written here.
+type Discovery struct {
+	Host  string
+	Path  string
+	Query string
+	Shape ListingShape
+	// Headers are static headers the vendor requires beyond the credential
+	// (Anthropic versions its API through one and rejects a request without
+	// it). The auth header itself comes from AuthHeaderName/Template.
+	Headers map[string]string
+}
+
+// RegionPlaceholder is replaced in Discovery.Host by the provider record's
+// configured region.
+const RegionPlaceholder = "<region>"
+
 // ExtraHeader names a single optional per-provider routing/config
 // header. Catalog declares N of these per provider type; the operator
 // fills any subset on the provider record (see Provider.ExtraValues).
@@ -245,8 +298,12 @@ var providers = []Provider{
 		AuthHeaderTemplate: "Bearer ${API_KEY}",
 		DefaultContentType: "application/json",
 		BrandColor:         "#10A37F",
-		ParserID:           "openai",
-		PricingSurfaces:    []string{"openai"},
+		Discovery: &Discovery{
+			Path:  "/v1/models",
+			Shape: ShapeOpenAIData,
+		},
+		ParserID:        "openai",
+		PricingSurfaces: []string{"openai"},
 		// Pricing + context windows cross-checked against LiteLLM's
 		// model_prices_and_context_window.json. Notable corrections from
 		// earlier values: o4-mini repriced from $4/$16 to $1.10/$4.40
@@ -284,8 +341,18 @@ var providers = []Provider{
 		AuthHeaderTemplate: "${API_KEY}",
 		DefaultContentType: "application/json",
 		BrandColor:         "#D97757",
-		ParserID:           "anthropic",
-		PricingSurfaces:    []string{"anthropic"},
+		Discovery: &Discovery{
+			Path: "/v1/models",
+			// The default page is short and a picker wants the whole
+			// catalogue in one call.
+			Query: "limit=1000",
+			Shape: ShapeOpenAIData,
+			// Anthropic versions its API through a header and refuses a
+			// request that omits it, listing included.
+			Headers: map[string]string{"anthropic-version": "2023-06-01"},
+		},
+		ParserID:        "anthropic",
+		PricingSurfaces: []string{"anthropic"},
 		// Per Anthropic's current model lineup. Pricing in USD per 1k
 		// tokens. Context windows: 4.6+ family is 1M; Haiku 4.5 stays at
 		// 200K. claude-3-7-sonnet and claude-3-5-haiku retired
@@ -296,6 +363,8 @@ var providers = []Provider{
 		// account to be on >= 30-day data retention or all requests
 		// 400.
 		Models: []Model{
+			{ID: "claude-opus-5", Label: "Claude Opus 5", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
+			{ID: "claude-sonnet-5", Label: "Claude Sonnet 5", InputPer1k: 0.003, OutputPer1k: 0.015, CacheReadPer1k: 0.0003, CacheCreationPer1k: 0.00375, ContextWindow: 1000000},
 			{ID: "claude-fable-5", Label: "Claude Fable 5", InputPer1k: 0.010, OutputPer1k: 0.050, CacheReadPer1k: 0.001, CacheCreationPer1k: 0.0125, ContextWindow: 1000000},
 			{ID: "claude-opus-4-8", Label: "Claude Opus 4.8", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
 			{ID: "claude-opus-4-7", Label: "Claude Opus 4.7", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
@@ -343,6 +412,22 @@ var providers = []Provider{
 		AuthHeaderTemplate: "Bearer ${API_KEY}",
 		DefaultContentType: "application/json",
 		BrandColor:         "#FF9900",
+		// Listings come from the CONTROL PLANE, not the runtime host in
+		// DefaultHost above: ListInferenceProfiles is not an operation
+		// bedrock-runtime implements, and answers <UnknownOperationException/>
+		// there. Inference has to go to the runtime host, so the two hosts
+		// genuinely differ and Discovery.Host carries the difference.
+		//
+		// Inference profiles rather than foundation models because the profile
+		// id is the invocable one: it carries the region prefix (eu., us.,
+		// global.) that AWS requires and that cannot be derived from the
+		// configured region — an eu-central-1 account legitimately holds
+		// global.* profiles.
+		Discovery: &Discovery{
+			Host:  "bedrock." + RegionPlaceholder + ".amazonaws.com",
+			Path:  "/inference-profiles",
+			Shape: ShapeBedrockInferenceProfiles,
+		},
 		// ParserID stays empty (path-style dispatch via IsBedrockPathStyle);
 		// the request parser meters these under the "bedrock" surface.
 		PricingSurfaces: []string{"bedrock"},
@@ -355,6 +440,8 @@ var providers = []Provider{
 		// Llama 3.3 70B entry kept unchanged — LiteLLM tracks only
 		// per-region Llama 3 entries; standalone 3.3 not yet listed.
 		Models: []Model{
+			{ID: "anthropic.claude-opus-5", Label: "Claude Opus 5 (Bedrock)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
+			{ID: "anthropic.claude-sonnet-5", Label: "Claude Sonnet 5 (Bedrock)", InputPer1k: 0.003, OutputPer1k: 0.015, CacheReadPer1k: 0.0003, CacheCreationPer1k: 0.00375, ContextWindow: 1000000},
 			{ID: "anthropic.claude-opus-4-8", Label: "Claude Opus 4.8 (Bedrock)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
 			{ID: "anthropic.claude-opus-4-7", Label: "Claude Opus 4.7 (Bedrock)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
 			{ID: "anthropic.claude-opus-4-6", Label: "Claude Opus 4.6 (Bedrock)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
@@ -391,6 +478,15 @@ var providers = []Provider{
 		AuthHeaderTemplate: "Bearer ${API_KEY}",
 		DefaultContentType: "application/json",
 		BrandColor:         "#4285F4",
+		// Only the v1beta1 publisher listing answers: the v1 form and the
+		// project-scoped form under BOTH versions return 404. That means the
+		// list is publisher-global — it cannot say which models this project
+		// has enabled — so it is offered as a suggestion beside the catalog
+		// rather than replacing it. See the discovery e2e for the probes.
+		Discovery: &Discovery{
+			Path:  "/v1beta1/publishers/anthropic/models",
+			Shape: ShapeVertexPublisherModels,
+		},
 		// ParserID stays empty (path-style dispatch via IsVertexPathStyle);
 		// Anthropic-on-Vertex requests are metered under the "anthropic"
 		// surface with the bare, unversioned model id.
@@ -406,6 +502,8 @@ var providers = []Provider{
 		// exists — the router denies unmeterable publishers rather than forward
 		// them uncounted.
 		Models: []Model{
+			{ID: "claude-opus-5", Label: "Claude Opus 5 (Vertex)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
+			{ID: "claude-sonnet-5", Label: "Claude Sonnet 5 (Vertex)", InputPer1k: 0.003, OutputPer1k: 0.015, CacheReadPer1k: 0.0003, CacheCreationPer1k: 0.00375, ContextWindow: 1000000},
 			{ID: "claude-fable-5", Label: "Claude Fable 5 (Vertex)", InputPer1k: 0.010, OutputPer1k: 0.050, CacheReadPer1k: 0.001, CacheCreationPer1k: 0.0125, ContextWindow: 1000000},
 			{ID: "claude-opus-4-8", Label: "Claude Opus 4.8 (Vertex)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},
 			{ID: "claude-opus-4-7", Label: "Claude Opus 4.7 (Vertex)", InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625, ContextWindow: 1000000},

management/internals/modules/agentnetwork/catalog/catalog_test.go

@@ -0,0 +1,36 @@
+package catalog
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// TestClaudeLineupSelectable pins the models Claude Code resolves to by
+// default. A model absent from the lineup can't be ticked on a provider
+// record, so llm_router denies it as not-routable and the operator has no
+// way to authorise the client's own default.
+func TestClaudeLineupSelectable(t *testing.T) {
+	for providerID, wanted := range map[string][]string{
+		"anthropic_api": {"claude-opus-5", "claude-sonnet-5", "claude-haiku-4-5"},
+		"bedrock_api":   {"anthropic.claude-opus-5", "anthropic.claude-sonnet-5", "anthropic.claude-haiku-4-5"},
+		"vertex_ai_api": {"claude-opus-5", "claude-sonnet-5", "claude-haiku-4-5"},
+	} {
+		provider, ok := Lookup(providerID)
+		require.True(t, ok, "catalog must define %s", providerID)
+
+		selectable := make(map[string]Model, len(provider.Models))
+		for _, m := range provider.Models {
+			selectable[m.ID] = m
+		}
+		for _, id := range wanted {
+			model, found := selectable[id]
+			require.True(t, found, "%s must offer %s", providerID, id)
+			assert.NotEmpty(t, model.Label, "%s/%s needs a label for the picker", providerID, id)
+			assert.Positive(t, model.InputPer1k, "%s/%s needs an input rate", providerID, id)
+			assert.Positive(t, model.OutputPer1k, "%s/%s needs an output rate", providerID, id)
+			assert.Positive(t, model.ContextWindow, "%s/%s needs a context window", providerID, id)
+		}
+	}
+}

management/internals/modules/agentnetwork/handlers/budget_handler_test.go

@@ -102,8 +102,8 @@ func TestSettingsHandler_GetExposesCollectionToggles(t *testing.T) {
 
 	require.NoError(t, f.store.SaveAgentNetworkSettings(context.Background(), &agentNetworkTypes.Settings{
 		AccountID:              testAccountID,
-		Cluster:                "eu.proxy.netbird.io",
-		Subdomain:              "violet",
+		Domain:                 "violet.eu.proxy.netbird.io",
+		ProxyAddress:           "eu.proxy.netbird.io",
 		EnableLogCollection:    true,
 		EnablePromptCollection: true,
 		RedactPii:              false,

management/internals/modules/agentnetwork/handlers/handlers_test.go

@@ -10,7 +10,7 @@ import (
 	"strings"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"

management/internals/modules/agentnetwork/handlers/model_discovery_handler_test.go

@@ -0,0 +1,178 @@
+package handlers
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/modeldiscovery"
+	nbcontext "github.com/netbirdio/netbird/management/server/context"
+	"github.com/netbirdio/netbird/shared/auth"
+	"github.com/netbirdio/netbird/shared/management/http/api"
+)
+
+// discoveryManagerStub records what the handler asked for and returns a canned
+// answer. The Manager interface is embedded rather than implemented: only the
+// one method is reachable from this handler, and a call to any other should
+// fail loudly rather than silently return a zero value.
+type discoveryManagerStub struct {
+	agentnetwork.Manager
+
+	gotReq      modeldiscovery.Request
+	gotRecordID string
+	models      []modeldiscovery.Model
+	err         error
+}
+
+func (s *discoveryManagerStub) DiscoverProviderModels(
+	_ context.Context, _, _ string, req modeldiscovery.Request, recordID string,
+) ([]modeldiscovery.Model, error) {
+	s.gotReq = req
+	s.gotRecordID = recordID
+	return s.models, s.err
+}
+
+// postDiscovery drives the handler with an authenticated request.
+func postDiscovery(t *testing.T, stub *discoveryManagerStub, body string) *httptest.ResponseRecorder {
+	t.Helper()
+	h := &handler{manager: stub}
+
+	req := httptest.NewRequest(http.MethodPost, "/agent-network/catalog/providers/models", strings.NewReader(body))
+	req = req.WithContext(nbcontext.SetUserAuthInContext(req.Context(), auth.UserAuth{
+		AccountId: "acc-1",
+		UserId:    "user-1",
+	}))
+
+	rec := httptest.NewRecorder()
+	h.discoverProviderModels(rec, req)
+	return rec
+}
+
+func TestDiscoverModelsReturnsTheVendorList(t *testing.T) {
+	stub := &discoveryManagerStub{models: []modeldiscovery.Model{
+		{ID: "eu.anthropic.claude-haiku-4-5-20251001-v1:0", Label: "EU Claude Haiku 4.5", PricingKnown: true},
+		{ID: "global.cohere.embed-v4:0", Label: "Global Cohere Embed v4"},
+		// A vendor that supplies no display name at all. Bedrock does for
+		// every profile, but the OpenAI listing carries none.
+		{ID: "gpt-4o-mini", PricingKnown: true},
+	}}
+
+	rec := postDiscovery(t, stub, `{
+		"catalog_provider_id":"bedrock_api",
+		"upstream_url":"https://bedrock-runtime.eu-central-1.amazonaws.com",
+		"api_key":"aws-bearer"
+	}`)
+	require.Equal(t, http.StatusOK, rec.Code, "body: %s", rec.Body.String())
+
+	var out api.AgentNetworkModelDiscoveryResponse
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &out))
+	require.Len(t, out.Models, 3)
+
+	assert.Equal(t, "eu.anthropic.claude-haiku-4-5-20251001-v1:0", out.Models[0].Id)
+	assert.True(t, out.Models[0].PricingKnown)
+	// An unpriced model must say so rather than arriving indistinguishable
+	// from a priced one: registering it silently would meter at zero.
+	assert.False(t, out.Models[1].PricingKnown)
+
+	require.NotNil(t, out.Models[0].Label, "the vendor supplied a display name")
+	assert.Equal(t, "EU Claude Haiku 4.5", *out.Models[0].Label)
+	// A vendor that supplies no name must omit the key rather than send an
+	// empty string: the dashboard falls back to the id on absence, and would
+	// render a blank row for "".
+	assert.Nil(t, out.Models[2].Label, "an absent label must not serialize")
+	assert.NotContains(t, rec.Body.String(), `"label":""`)
+
+	assert.Equal(t, "bedrock_api", stub.gotReq.CatalogID)
+	assert.Equal(t, "aws-bearer", stub.gotReq.APIKey)
+	// The upstream is what the region is read back out of for Bedrock, so
+	// losing it here would break discovery for every regional provider.
+	assert.Equal(t, "https://bedrock-runtime.eu-central-1.amazonaws.com", stub.gotReq.UpstreamURL)
+	assert.Empty(t, stub.gotRecordID)
+}
+
+func TestDiscoverModelsUsesAStoredRecordWithoutAKey(t *testing.T) {
+	stub := &discoveryManagerStub{}
+
+	rec := postDiscovery(t, stub, `{"catalog_provider_id":"openai_api","provider_id":"prov-42"}`)
+	require.Equal(t, http.StatusOK, rec.Code, "body: %s", rec.Body.String())
+
+	// The dashboard refreshes a saved provider's list without ever holding
+	// the credential, so the record id has to reach the manager.
+	assert.Equal(t, "prov-42", stub.gotRecordID)
+	assert.Empty(t, stub.gotReq.APIKey)
+}
+
+// TestDiscoverModelsRefusesMixedCredentials covers the case where a caller
+// names a saved provider AND supplies a key. Accepting it would run an
+// arbitrary credential under the identity of a record the caller may only be
+// permitted to read.
+func TestDiscoverModelsRefusesMixedCredentials(t *testing.T) {
+	stub := &discoveryManagerStub{}
+
+	rec := postDiscovery(t, stub, `{
+		"catalog_provider_id":"openai_api",
+		"provider_id":"prov-42",
+		"api_key":"sk-attacker"
+	}`)
+	assert.Equal(t, http.StatusBadRequest, rec.Code)
+	assert.Empty(t, stub.gotRecordID, "the request must be refused before it reaches the manager")
+}
+
+// TestDiscoverModelsReportsNoDiscoveryDistinctly matters because the caller
+// falls back to the catalog's own model list on this outcome. Collapsing it
+// into a generic 500 would turn "this provider has no listing endpoint" into
+// "something went wrong", and the form would show an error instead of a list.
+func TestDiscoverModelsReportsNoDiscoveryDistinctly(t *testing.T) {
+	stub := &discoveryManagerStub{err: modeldiscovery.ErrNoDiscovery}
+
+	rec := postDiscovery(t, stub, `{"catalog_provider_id":"litellm_proxy","upstream_url":"https://gw.example.com","api_key":"sk"}`)
+	assert.Equal(t, http.StatusUnprocessableEntity, rec.Code)
+}
+
+// TestDiscoverModelsTrimsTheCatalogID pins that the id the emptiness check
+// accepts is the id the manager receives. A padded value that clears the check
+// but reaches the catalog untrimmed misses the lookup, and the operator is told
+// their provider does not exist.
+func TestDiscoverModelsTrimsTheCatalogID(t *testing.T) {
+	stub := &discoveryManagerStub{}
+
+	rec := postDiscovery(t, stub, `{"catalog_provider_id":"  openai_api  ","api_key":"sk"}`)
+	require.Equal(t, http.StatusOK, rec.Code, "body: %s", rec.Body.String())
+	assert.Equal(t, "openai_api", stub.gotReq.CatalogID)
+}
+
+// TestDiscoverModelsReportsCallerInputAsBadRequest covers the other half of the
+// error mapping. These failures are all reachable from a well-formed request
+// with a bad field value, so answering 500 both misinforms the operator and
+// puts their typo into the server's error rate.
+func TestDiscoverModelsReportsCallerInputAsBadRequest(t *testing.T) {
+	stub := &discoveryManagerStub{
+		err: fmt.Errorf("%w: unknown catalog provider %q", modeldiscovery.ErrInvalidRequest, "nope"),
+	}
+
+	rec := postDiscovery(t, stub, `{"catalog_provider_id":"nope","api_key":"sk"}`)
+	assert.Equal(t, http.StatusBadRequest, rec.Code)
+	assert.Contains(t, rec.Body.String(), "unknown catalog provider")
+}
+
+func TestDiscoverModelsRejectsMalformedRequests(t *testing.T) {
+	for name, body := range map[string]string{
+		"not json":               `{`,
+		"no catalog provider":    `{"api_key":"sk"}`,
+		"blank catalog provider": `{"catalog_provider_id":"   ","api_key":"sk"}`,
+	} {
+		t.Run(name, func(t *testing.T) {
+			stub := &discoveryManagerStub{}
+			rec := postDiscovery(t, stub, body)
+			assert.Equal(t, http.StatusBadRequest, rec.Code)
+		})
+	}
+}

management/internals/modules/agentnetwork/handlers/providers_handler.go

@@ -7,6 +7,7 @@ package handlers
 
 import (
 	"encoding/json"
+	"errors"
 	"math"
 	"net/http"
 	"net/url"
@@ -16,6 +17,7 @@ import (
 
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/catalog"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/modeldiscovery"
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/pricing"
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
 	nbcontext "github.com/netbirdio/netbird/management/server/context"
@@ -32,6 +34,7 @@ type handler struct {
 func RegisterEndpoints(manager agentnetwork.Manager, router *mux.Router) {
 	h := &handler{manager: manager}
 	router.HandleFunc("/agent-network/catalog/providers", h.getCatalogProviders).Methods("GET", "OPTIONS")
+	router.HandleFunc("/agent-network/catalog/providers/models", h.discoverProviderModels).Methods("POST", "OPTIONS")
 	router.HandleFunc("/agent-network/providers", h.getAllProviders).Methods("GET", "OPTIONS")
 	router.HandleFunc("/agent-network/providers", h.createProvider).Methods("POST", "OPTIONS")
 	router.HandleFunc("/agent-network/providers/{providerId}", h.getProvider).Methods("GET", "OPTIONS")
@@ -61,6 +64,98 @@ func (h *handler) getCatalogProviders(w http.ResponseWriter, r *http.Request) {
 	util.WriteJSONObject(r.Context(), w, out)
 }
 
+// discoverProviderModels asks the vendor which models the operator's own
+// credential can reach, so the provider form can offer a live list rather than
+// only the static catalog.
+func (h *handler) discoverProviderModels(w http.ResponseWriter, r *http.Request) {
+	userAuth, err := nbcontext.GetUserAuthFromContext(r.Context())
+	if err != nil {
+		util.WriteError(r.Context(), err, w)
+		return
+	}
+
+	var body api.AgentNetworkModelDiscoveryRequest
+	if err := json.NewDecoder(r.Body).Decode(&body); err != nil {
+		util.WriteErrorResponse("invalid json", http.StatusBadRequest, w)
+		return
+	}
+	// Trimmed once and carried, not trimmed for the emptiness test and then
+	// discarded: a padded " openai_api " would clear the check here and miss
+	// the catalog lookup, reporting the provider as unknown.
+	catalogID := strings.TrimSpace(body.CatalogProviderId)
+	if catalogID == "" {
+		util.WriteErrorResponse("catalog_provider_id is required", http.StatusBadRequest, w)
+		return
+	}
+
+	recordID := strValue(body.ProviderId)
+	req := modeldiscovery.Request{
+		CatalogID:   catalogID,
+		UpstreamURL: strValue(body.UpstreamUrl),
+		APIKey:      strValue(body.ApiKey),
+	}
+	// One source of credential or the other, never a mix: taking a key from
+	// the request while addressing a saved record would let a caller run an
+	// arbitrary credential against a provider they can only read.
+	if recordID != "" && req.APIKey != "" {
+		util.WriteErrorResponse("provide either provider_id or api_key, not both", http.StatusBadRequest, w)
+		return
+	}
+
+	models, err := h.manager.DiscoverProviderModels(r.Context(), userAuth.AccountId, userAuth.UserId, req, recordID)
+	if err != nil {
+		// A provider with no listing endpoint is a fact about the catalog
+		// entry, not a failure: the caller falls back to the catalog's own
+		// models, so it must be able to tell the two apart.
+		if errors.Is(err, modeldiscovery.ErrNoDiscovery) {
+			util.WriteErrorResponse(err.Error(), http.StatusUnprocessableEntity, w)
+			return
+		}
+		// An unknown provider, an unusable upstream, a missing region or a
+		// missing key are all things the caller sent, reachable from a
+		// well-formed request. Reporting them as 500 tells the operator the
+		// server broke and buries genuine faults in the error rate.
+		if errors.Is(err, modeldiscovery.ErrInvalidRequest) {
+			util.WriteErrorResponse(err.Error(), http.StatusBadRequest, w)
+			return
+		}
+		util.WriteError(r.Context(), err, w)
+		return
+	}
+
+	out := api.AgentNetworkModelDiscoveryResponse{Models: make([]api.AgentNetworkDiscoveredModel, 0, len(models))}
+	for _, m := range models {
+		entry := api.AgentNetworkDiscoveredModel{
+			Id:           m.ID,
+			PricingKnown: m.PricingKnown,
+			// Sent even when zero: the form prefills every discovered model as
+			// an editable row, and an unpriced one is shown at zero and flagged
+			// rather than left out.
+			InputPer1k:  m.InputPer1k,
+			OutputPer1k: m.OutputPer1k,
+			// Cache rates stay absent when unset, matching the catalog
+			// response — a zero would read as "free", not "not applicable".
+			CachedInputPer1k:   positiveRatePtr(m.CachedInputPer1k),
+			CacheReadPer1k:     positiveRatePtr(m.CacheReadPer1k),
+			CacheCreationPer1k: positiveRatePtr(m.CacheCreationPer1k),
+		}
+		if m.Label != "" {
+			label := m.Label
+			entry.Label = &label
+		}
+		out.Models = append(out.Models, entry)
+	}
+	util.WriteJSONObject(r.Context(), w, out)
+}
+
+// strValue reads an optional string field, treating absent as empty.
+func strValue(v *string) string {
+	if v == nil {
+		return ""
+	}
+	return strings.TrimSpace(*v)
+}
+
 // applyDefaultPricing overwrites the catalog response's model rates with
 // the LIVE default pricing table, which may differ from the compiled-in
 // catalog rates when the operator provides a defaults_llm_pricing.yaml.
@@ -155,12 +250,7 @@ func (h *handler) createProvider(w http.ResponseWriter, r *http.Request) {
 	provider := types.NewProvider(userAuth.AccountId)
 	provider.FromAPIRequest(&req)
 
-	bootstrapCluster := ""
-	if req.BootstrapCluster != nil {
-		bootstrapCluster = *req.BootstrapCluster
-	}
-
-	created, err := h.manager.CreateProvider(r.Context(), userAuth.UserId, provider, bootstrapCluster)
+	created, err := h.manager.CreateProvider(r.Context(), userAuth.UserId, provider)
 	if err != nil {
 		util.WriteError(r.Context(), err, w)
 		return

management/internals/modules/agentnetwork/handlers/settings_handler.go

@@ -12,13 +12,55 @@ import (
 	"github.com/netbirdio/netbird/shared/management/http/util"
 )
 
-// addSettingsEndpoints registers the Agent Network settings routes. The
-// settings row is bootstrapped server-side on first provider create or on the
-// first PUT carrying a cluster; GET reads it and PUT applies a partial update
-// of the mutable collection toggles (cluster/subdomain stay immutable).
+// addSettingsEndpoints registers the Agent Network settings routes. POST
+// bootstraps the settings row, assigning the account's immutable endpoint;
+// GET reads it (defaults with an empty endpoint before bootstrap); PUT
+// carries every field, replacing the mutable collection toggles and rejecting
+// any change to the identity fields; DELETE removes the row — guarded so it
+// stays a bootstrap-repair operation — releasing the endpoint for a fresh
+// bootstrap.
 func (h *handler) addSettingsEndpoints(router *mux.Router) {
 	router.HandleFunc("/agent-network/settings", h.getSettings).Methods("GET", "OPTIONS")
+	router.HandleFunc("/agent-network/settings", h.createSettings).Methods("POST", "OPTIONS")
 	router.HandleFunc("/agent-network/settings", h.updateSettings).Methods("PUT", "OPTIONS")
+	router.HandleFunc("/agent-network/settings", h.deleteSettings).Methods("DELETE", "OPTIONS")
+}
+
+// createSettings bootstraps the account's settings row. Exactly one of
+// proxy_address (labeled endpoint; the server allocates the label) and
+// endpoint (self-addressed, claimed verbatim) must be provided; optional
+// collection toggles ride along with defaults for omitted fields.
+func (h *handler) createSettings(w http.ResponseWriter, r *http.Request) {
+	userAuth, err := nbcontext.GetUserAuthFromContext(r.Context())
+	if err != nil {
+		util.WriteError(r.Context(), err, w)
+		return
+	}
+
+	var req api.AgentNetworkSettingsCreateRequest
+	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+		util.WriteErrorResponse("couldn't parse JSON request", http.StatusBadRequest, w)
+		return
+	}
+
+	settings := types.DefaultSettings(userAuth.AccountId)
+	settings.FromAPICreateRequest(&req)
+
+	proxyAddress := ""
+	if req.ProxyAddress != nil {
+		proxyAddress = *req.ProxyAddress
+	}
+	endpoint := ""
+	if req.Endpoint != nil {
+		endpoint = *req.Endpoint
+	}
+
+	created, err := h.manager.CreateSettings(r.Context(), userAuth.UserId, settings, proxyAddress, endpoint)
+	if err != nil {
+		util.WriteError(r.Context(), err, w)
+		return
+	}
+	util.WriteJSONObject(r.Context(), w, created.ToAPIResponse())
 }
 
 // updateSettings replaces the mutable settings fields on the account's row.
@@ -48,6 +90,24 @@ func (h *handler) updateSettings(w http.ResponseWriter, r *http.Request) {
 	util.WriteJSONObject(r.Context(), w, updated.ToAPIResponse())
 }
 
+// deleteSettings removes the account's settings row, releasing the endpoint.
+// The manager refuses (412) while providers exist or a proxy is actively
+// serving the endpoint; a later POST bootstraps fresh, allocating a new
+// endpoint.
+func (h *handler) deleteSettings(w http.ResponseWriter, r *http.Request) {
+	userAuth, err := nbcontext.GetUserAuthFromContext(r.Context())
+	if err != nil {
+		util.WriteError(r.Context(), err, w)
+		return
+	}
+
+	if err := h.manager.DeleteSettings(r.Context(), userAuth.AccountId, userAuth.UserId); err != nil {
+		util.WriteError(r.Context(), err, w)
+		return
+	}
+	util.WriteJSONObject(r.Context(), w, util.EmptyObject{})
+}
+
 // getSettings returns the account's agent-network settings. Accounts that
 // haven't been bootstrapped yet read as the defaults with an empty cluster,
 // subdomain and endpoint; the manager synthesises that view.

management/internals/modules/agentnetwork/handlers/settings_handler_test.go

@@ -1,20 +1,25 @@
 package handlers
 
 import (
+	"context"
 	"encoding/json"
+	"fmt"
 	"net/http"
+	"strings"
 	"testing"
+	"time"
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
+	rpproxy "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
 	"github.com/netbirdio/netbird/shared/management/http/api"
 )
 
 // TestSettingsHandler_GetUnbootstrappedReturnsDefaults pins the settings-read
 // convention shared with the account and DNS settings endpoints: settings
 // always read as a JSON object. Before bootstrap that object carries the
-// defaults with an empty cluster/subdomain/endpoint (the "not bootstrapped"
+// defaults with an empty endpoint/proxy_address (the "not bootstrapped"
 // signal) and no timestamps — never a 404 and never the legacy null body.
 func TestSettingsHandler_GetUnbootstrappedReturnsDefaults(t *testing.T) {
 	f := newAgentNetworkHandlerFixture(t)
@@ -27,9 +32,9 @@ func TestSettingsHandler_GetUnbootstrappedReturnsDefaults(t *testing.T) {
 
 	var got api.AgentNetworkSettings
 	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
-	assert.Empty(t, got.Cluster, "cluster must be empty until bootstrapped")
-	assert.Empty(t, got.Subdomain, "subdomain must be empty until bootstrapped")
-	assert.Empty(t, got.Endpoint, "endpoint must be empty until bootstrapped, not a bare dot")
+	assert.Empty(t, got.Endpoint, "endpoint must be empty until bootstrapped")
+	assert.Empty(t, got.ProxyAddress, "proxy address must be empty until bootstrapped")
+	assert.False(t, got.Dedicated, "an unbootstrapped account has no serving shape")
 	assert.True(t, got.EnableLogCollection, "defaults must show log collection on, matching bootstrap")
 	assert.False(t, got.EnablePromptCollection, "defaults must show prompt collection off")
 	assert.False(t, got.RedactPii, "defaults must show redaction off")
@@ -39,62 +44,149 @@ func TestSettingsHandler_GetUnbootstrappedReturnsDefaults(t *testing.T) {
 	assert.Nil(t, got.UpdatedAt, "no timestamps before a row exists")
 }
 
-// TestSettingsHandler_PutBootstrapsWithCluster covers the settings-first
-// bootstrap path: a PUT carrying a cluster on an unbootstrapped account
-// creates the row (cluster pinned, subdomain assigned) and applies the
-// mutable fields from the same request.
-func TestSettingsHandler_PutBootstrapsWithCluster(t *testing.T) {
+// TestSettingsHandler_PostBootstrapsLabeled covers the labeled bootstrap
+// shape: a POST carrying a proxy_address allocates a label beneath it, so the
+// endpoint hangs one label under the shared cluster's address and the pin is
+// not dedicated. Toggles riding along apply; omitted ones keep defaults.
+func TestSettingsHandler_PostBootstrapsLabeled(t *testing.T) {
 	f := newAgentNetworkHandlerFixture(t)
 
-	rec := f.do(t, http.MethodPut, "/agent-network/settings",
-		`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": true, "redact_pii": false, "access_log_retention_days": 30}`)
-	require.Equal(t, http.StatusOK, rec.Code, "bootstrap PUT must succeed: %s", rec.Body.String())
+	rec := f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true, "access_log_retention_days": 14}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
 
 	var got api.AgentNetworkSettings
 	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
-	assert.Equal(t, "eu.proxy.netbird.io", got.Cluster, "cluster must be pinned from the request")
-	assert.NotEmpty(t, got.Subdomain, "subdomain must be assigned at bootstrap")
-	assert.Equal(t, got.Subdomain+".eu.proxy.netbird.io", got.Endpoint, "endpoint must combine subdomain and cluster")
-	assert.True(t, got.EnableLogCollection, "toggle from the bootstrap request must apply")
+	assert.Equal(t, "eu.proxy.netbird.io", got.ProxyAddress, "proxy address must be pinned from the request")
+	require.NotEmpty(t, got.Endpoint, "endpoint must be allocated at bootstrap")
+	assert.True(t, strings.HasSuffix(got.Endpoint, ".eu.proxy.netbird.io"),
+		"labeled endpoint must hang off the proxy address: %s", got.Endpoint)
+	label := strings.TrimSuffix(got.Endpoint, ".eu.proxy.netbird.io")
+	assert.NotContains(t, label, ".", "the allocated label must be a single DNS label: %s", label)
+	assert.False(t, got.Dedicated, "a labeled pin is not dedicated")
+	assert.True(t, got.EnableLogCollection, "omitted toggle must keep its default")
 	assert.True(t, got.EnablePromptCollection, "toggle from the bootstrap request must apply")
 	require.NotNil(t, got.AccessLogRetentionDays)
-	assert.Equal(t, 30, *got.AccessLogRetentionDays, "retention from the bootstrap request must apply")
+	assert.Equal(t, 14, *got.AccessLogRetentionDays, "retention from the bootstrap request must apply")
+	assert.NotNil(t, got.CreatedAt, "a persisted row carries timestamps")
 
 	// The row is now readable via GET.
 	rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
 	require.Equal(t, http.StatusOK, rec.Code, "GET after bootstrap must succeed")
+	var read api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &read))
+	assert.Equal(t, got.Endpoint, read.Endpoint, "GET must return the bootstrapped endpoint")
 }
 
-// TestSettingsHandler_PutWithoutClusterOnUnbootstrapped pins that a PUT
-// without a cluster cannot conjure a settings row out of nothing — there is
-// no cluster to pin — and surfaces as 404 like the GET.
-func TestSettingsHandler_PutWithoutClusterOnUnbootstrapped(t *testing.T) {
+// TestSettingsHandler_PostBootstrapsSelfAddressed covers the dedicated shape:
+// a POST carrying an endpoint claims the hostname verbatim, the proxy address
+// equals it, and the pin reads as dedicated. The claim is legitimate before
+// any proxy declares the address (address-first).
+func TestSettingsHandler_PostBootstrapsSelfAddressed(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"endpoint": "Brave-Otter.Gateway.Example.com"}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
+
+	var got api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
+	assert.Equal(t, "brave-otter.gateway.example.com", got.Endpoint,
+		"endpoint must be claimed verbatim, lowercased")
+	assert.Equal(t, got.Endpoint, got.ProxyAddress, "self-addressed: the proxy address is the endpoint")
+	assert.True(t, got.Dedicated, "a self-addressed pin is dedicated")
+	assert.True(t, got.EnableLogCollection, "omitted toggles must keep their defaults")
+}
+
+// TestSettingsHandler_PostRequiresExactlyOneIdentityField pins the request
+// contract: proxy_address and endpoint are mutually exclusive and one is
+// required — both or neither is a validation error, not a guess.
+func TestSettingsHandler_PostRequiresExactlyOneIdentityField(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodPost, "/agent-network/settings", `{}`)
+	assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
+		"empty POST must be rejected: got %d body=%s", rec.Code, rec.Body.String())
+
+	rec = f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"proxy_address": "eu.proxy.netbird.io", "endpoint": "brave-otter.gateway.example.com"}`)
+	assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
+		"POST with both identity fields must be rejected: got %d body=%s", rec.Code, rec.Body.String())
+}
+
+// TestSettingsHandler_PostRejectsMalformedHostnames pins per-write input
+// validation: shapes canonicalization cannot repair — trailing dots, embedded
+// whitespace, empty labels — are rejected with a validation error instead of
+// landing in an immutable column.
+func TestSettingsHandler_PostRejectsMalformedHostnames(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	for name, body := range map[string]string{
+		"trailing dot":     `{"endpoint": "gateway.example.com."}`,
+		"leading dot":      `{"endpoint": ".gateway.example.com"}`,
+		"inner whitespace": `{"endpoint": "gate way.example.com"}`,
+		"empty label":      `{"proxy_address": "eu..proxy.netbird.io"}`,
+	} {
+		rec := f.do(t, http.MethodPost, "/agent-network/settings", body)
+		assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
+			"%s must be rejected: got %d body=%s", name, rec.Code, rec.Body.String())
+	}
+}
+
+// TestSettingsHandler_PostConflictsOnSecondBootstrap pins that bootstrap is a
+// one-time create: a second POST returns 409 and leaves the row untouched.
+func TestSettingsHandler_PostConflictsOnSecondBootstrap(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "eu.proxy.netbird.io"}`)
+	require.Equal(t, http.StatusOK, rec.Code, "first bootstrap must succeed: %s", rec.Body.String())
+	var first api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &first))
+
+	rec = f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "us.proxy.netbird.io"}`)
+	assert.Equal(t, http.StatusConflict, rec.Code,
+		"second bootstrap must 409: got %d body=%s", rec.Code, rec.Body.String())
+
+	rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
+	require.Equal(t, http.StatusOK, rec.Code)
+	var got api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
+	assert.Equal(t, first.Endpoint, got.Endpoint, "the original endpoint must survive the rejected bootstrap")
+	assert.Equal(t, first.ProxyAddress, got.ProxyAddress, "the original proxy address must survive")
+}
+
+// TestSettingsHandler_PutBeforeBootstrapIs404 pins that a PUT cannot conjure a
+// settings row out of nothing — bootstrap is the explicit POST — and the
+// error points the caller there.
+func TestSettingsHandler_PutBeforeBootstrapIs404(t *testing.T) {
 	f := newAgentNetworkHandlerFixture(t)
 
 	rec := f.do(t, http.MethodPut, "/agent-network/settings",
 		`{"enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false}`)
 	assert.Equal(t, http.StatusNotFound, rec.Code,
-		"cluster-less PUT on an unbootstrapped account must 404: got %d body=%s", rec.Code, rec.Body.String())
-	assert.Contains(t, rec.Body.String(), "cluster",
-		"the error must point the caller at the bootstrap paths: %s", rec.Body.String())
+		"PUT on an unbootstrapped account must 404: got %d body=%s", rec.Code, rec.Body.String())
+	assert.Contains(t, rec.Body.String(), "/api/agent-network/settings",
+		"the error must point the caller at the bootstrap POST: %s", rec.Body.String())
 }
 
 // TestSettingsHandler_PutReplacesMutableFields pins the update contract shared
-// with the other PUT endpoints: the request replaces every mutable field, so a
-// toggle absent from the JSON lands as its zero value rather than being
-// preserved. Cluster and subdomain survive untouched.
+// with the other PUT endpoints: the request carries every field, replacing the
+// mutable ones. The identity fields ride along as a required echo of the
+// assigned values — compared, never written — so the endpoint and proxy
+// address survive every accepted update.
 func TestSettingsHandler_PutReplacesMutableFields(t *testing.T) {
 	f := newAgentNetworkHandlerFixture(t)
 
-	rec := f.do(t, http.MethodPut, "/agent-network/settings",
-		`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": true, "redact_pii": true, "access_log_retention_days": 14}`)
-	require.Equal(t, http.StatusOK, rec.Code, "bootstrap PUT must succeed: %s", rec.Body.String())
+	rec := f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true, "redact_pii": true, "access_log_retention_days": 14}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
 
 	var before api.AgentNetworkSettings
 	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
 
-	rec = f.do(t, http.MethodPut, "/agent-network/settings",
-		`{"enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`)
+	rec = f.do(t, http.MethodPut, "/agent-network/settings", fmt.Sprintf(
+		`{"endpoint": %q, "proxy_address": %q, "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 7}`,
+		before.Endpoint, before.ProxyAddress))
 	require.Equal(t, http.StatusOK, rec.Code, "update PUT must succeed: %s", rec.Body.String())
 
 	var got api.AgentNetworkSettings
@@ -103,35 +195,201 @@ func TestSettingsHandler_PutReplacesMutableFields(t *testing.T) {
 	assert.False(t, got.EnablePromptCollection, "sent toggle must apply")
 	assert.False(t, got.RedactPii, "sent toggle must apply")
 	require.NotNil(t, got.AccessLogRetentionDays)
-	assert.Equal(t, 0, *got.AccessLogRetentionDays,
-		"retention absent from the request must land as the zero value — PUT replaces all mutable fields")
-	assert.Equal(t, before.Cluster, got.Cluster, "cluster must survive updates untouched")
-	assert.Equal(t, before.Subdomain, got.Subdomain, "subdomain must survive updates untouched")
+	assert.Equal(t, 7, *got.AccessLogRetentionDays, "sent retention must apply")
+	assert.Equal(t, before.Endpoint, got.Endpoint, "endpoint must survive updates untouched")
+	assert.Equal(t, before.ProxyAddress, got.ProxyAddress, "proxy address must survive updates untouched")
 }
 
-// TestSettingsHandler_PutRejectsClusterChange pins cluster immutability: once
-// assigned, a differing cluster is rejected as a validation error instead of
-// being silently ignored, so callers never observe a value other than the one
-// they sent. Echoing the assigned cluster back stays valid, which lets
-// declarative clients send their full desired state idempotently.
-func TestSettingsHandler_PutRejectsClusterChange(t *testing.T) {
+// TestSettingsHandler_PutRejectsChangedIdentity pins the immutability contract:
+// the PUT carries the identity fields like every other field, but they are an
+// echo — a request carrying a different endpoint or proxy address is rejected
+// as a validation error and the row is left untouched. The comparison is
+// lenient about casing (the stored values are normalized lowercase), so a
+// client replaying a GET response with different casing is not rejected.
+func TestSettingsHandler_PutRejectsChangedIdentity(t *testing.T) {
 	f := newAgentNetworkHandlerFixture(t)
 
-	rec := f.do(t, http.MethodPut, "/agent-network/settings",
-		`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`)
-	require.Equal(t, http.StatusOK, rec.Code, "bootstrap PUT must succeed: %s", rec.Body.String())
+	rec := f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
+	var before api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
 
-	rec = f.do(t, http.MethodPut, "/agent-network/settings",
-		`{"cluster": "us.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`)
-	assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
-		"cluster change must be rejected as a validation error: got %d body=%s", rec.Code, rec.Body.String())
+	for name, body := range map[string]string{
+		"changed endpoint": fmt.Sprintf(
+			`{"endpoint": "other.gateway.example.com", "proxy_address": %q, "enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 30}`,
+			before.ProxyAddress),
+		"changed proxy_address": fmt.Sprintf(
+			`{"endpoint": %q, "proxy_address": "us.proxy.netbird.io", "enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 30}`,
+			before.Endpoint),
+		"omitted identity": `{"enable_log_collection": false, "enable_prompt_collection": false, "redact_pii": false, "access_log_retention_days": 30}`,
+	} {
+		rec = f.do(t, http.MethodPut, "/agent-network/settings", body)
+		assert.Equal(t, http.StatusUnprocessableEntity, rec.Code,
+			"%s must be rejected: got %d body=%s", name, rec.Code, rec.Body.String())
+	}
 
-	rec = f.do(t, http.MethodPut, "/agent-network/settings",
-		`{"cluster": "eu.proxy.netbird.io", "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": true}`)
-	require.Equal(t, http.StatusOK, rec.Code, "echoing the assigned cluster must stay valid: %s", rec.Body.String())
+	// The rejected updates must not have applied anything — toggles included.
+	rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
+	require.Equal(t, http.StatusOK, rec.Code)
+	var got api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
+	assert.Equal(t, before.Endpoint, got.Endpoint, "rejected PUT must not change the endpoint")
+	assert.Equal(t, before.ProxyAddress, got.ProxyAddress, "rejected PUT must not change the proxy address")
+	assert.True(t, got.EnablePromptCollection, "rejected PUT must not apply its toggles")
+
+	// An uppercased echo of the assigned values still names the same host and
+	// must be accepted.
+	rec = f.do(t, http.MethodPut, "/agent-network/settings", fmt.Sprintf(
+		`{"endpoint": %q, "proxy_address": %q, "enable_log_collection": true, "enable_prompt_collection": true, "redact_pii": false, "access_log_retention_days": 30}`,
+		strings.ToUpper(before.Endpoint), strings.ToUpper(before.ProxyAddress)))
+	assert.Equal(t, http.StatusOK, rec.Code,
+		"an uppercased identity echo must be accepted: got %d body=%s", rec.Code, rec.Body.String())
+}
+
+// TestSettingsHandler_PutOmittedRetentionLandsAsZero documents a residual the
+// required-ness of access_log_retention_days does not remove. Marking the field
+// required changes the generated client type from *int to int, so a generated
+// client cannot omit it — but nothing validates OpenAPI required-ness at
+// runtime, so a hand-rolled body without the field still decodes as 0, which
+// the API documents as "keep indefinitely".
+//
+// That is the same latitude the three booleans already have, so it is left
+// consistent rather than special-cased. This test exists to make the gap
+// explicit: if request validation is ever added, this expectation is what
+// changes.
+func TestSettingsHandler_PutOmittedRetentionLandsAsZero(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"proxy_address": "eu.proxy.netbird.io", "access_log_retention_days": 14}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
+	var before api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
+
+	rec = f.do(t, http.MethodPut, "/agent-network/settings", fmt.Sprintf(
+		`{"endpoint": %q, "proxy_address": %q, "enable_log_collection": true, "enable_prompt_collection": false, "redact_pii": false}`,
+		before.Endpoint, before.ProxyAddress))
+	require.Equal(t, http.StatusOK, rec.Code, "update PUT must succeed: %s", rec.Body.String())
 
 	var got api.AgentNetworkSettings
 	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
-	assert.Equal(t, "eu.proxy.netbird.io", got.Cluster, "cluster must be unchanged")
-	assert.True(t, got.RedactPii, "toggle sent alongside the echoed cluster must apply")
+	require.NotNil(t, got.AccessLogRetentionDays)
+	assert.Equal(t, 0, *got.AccessLogRetentionDays,
+		"a non-conforming body that omits retention still replaces it with the zero value")
+}
+
+// TestSettingsHandler_DeleteBeforeBootstrapIs404 pins that DELETE on an
+// account with no settings row is a 404, mirroring the PUT.
+func TestSettingsHandler_DeleteBeforeBootstrapIs404(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodDelete, "/agent-network/settings", "")
+	assert.Equal(t, http.StatusNotFound, rec.Code,
+		"DELETE on an unbootstrapped account must 404: got %d body=%s", rec.Code, rec.Body.String())
+}
+
+// TestSettingsHandler_DeleteBlockedByProviders pins the first delete guard:
+// while any provider exists for the account, the delete is refused with 412
+// and the row survives. Providers route through the endpoint — the guard
+// keeps DELETE a bootstrap-repair operation rather than a way to abandon a
+// configured gateway.
+func TestSettingsHandler_DeleteBlockedByProviders(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "eu.proxy.netbird.io"}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
+	var before api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &before))
+
+	f.seedProvider(t, "prov-guard")
+
+	rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
+	assert.Equal(t, http.StatusPreconditionFailed, rec.Code,
+		"delete with a provider present must be refused: got %d body=%s", rec.Code, rec.Body.String())
+
+	rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
+	require.Equal(t, http.StatusOK, rec.Code)
+	var got api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &got))
+	assert.Equal(t, before.Endpoint, got.Endpoint, "the refused delete must leave the row intact")
+}
+
+// TestSettingsHandler_DeleteBlockedByActiveProxy pins the second delete
+// guard: while a proxy is actively serving the endpoint — an active proxy
+// row declaring the endpoint hostname as its cluster address, the dedicated
+// shape — the delete is refused with 412. A proxy that has disconnected no
+// longer blocks: the guard is about a live serving path, not history.
+//
+// The proxy declares its address with mixed casing on purpose: Connect
+// stores the declared address verbatim while the settings row is normalized
+// lowercase, and hostnames are case-insensitive, so the guard must match
+// across the casing difference rather than be sidestepped by it.
+func TestSettingsHandler_DeleteBlockedByActiveProxy(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	const endpoint = "gw.dedicated.example.com"
+	rec := f.do(t, http.MethodPost, "/agent-network/settings", fmt.Sprintf(`{"endpoint": %q}`, endpoint))
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
+
+	now := time.Now()
+	accountID := testAccountID
+	proxyRow := &rpproxy.Proxy{
+		ID:             "proxy-guard",
+		SessionID:      "sess-1",
+		ClusterAddress: "GW.Dedicated.Example.Com",
+		AccountID:      &accountID,
+		LastSeen:       now,
+		ConnectedAt:    &now,
+		Status:         rpproxy.StatusConnected,
+	}
+	require.NoError(t, f.store.SaveProxy(context.Background(), proxyRow))
+
+	rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
+	assert.Equal(t, http.StatusPreconditionFailed, rec.Code,
+		"delete with an active proxy at the endpoint must be refused: got %d body=%s", rec.Code, rec.Body.String())
+
+	// Once the proxy disconnects it no longer serves the endpoint, so the
+	// delete goes through.
+	require.NoError(t, f.store.DisconnectProxy(context.Background(), proxyRow.ID, proxyRow.SessionID))
+	rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
+	assert.Equal(t, http.StatusOK, rec.Code,
+		"delete after the proxy disconnected must succeed: got %d body=%s", rec.Code, rec.Body.String())
+}
+
+// TestSettingsHandler_DeleteReleasesEndpointForFreshBootstrap pins the
+// full-reset semantic that gives replace-on-change clients (e.g. Terraform's
+// RequiresReplace) a real path: with both guards clear the delete succeeds,
+// the account reads as the defaults again, and a fresh bootstrap draws a
+// fresh label. The released hostname is not reserved — a fresh draw may even
+// legitimately re-pick it — so the assertions check the new row's shape, not
+// that the label differs.
+func TestSettingsHandler_DeleteReleasesEndpointForFreshBootstrap(t *testing.T) {
+	f := newAgentNetworkHandlerFixture(t)
+
+	rec := f.do(t, http.MethodPost, "/agent-network/settings",
+		`{"proxy_address": "eu.proxy.netbird.io", "enable_prompt_collection": true}`)
+	require.Equal(t, http.StatusOK, rec.Code, "bootstrap POST must succeed: %s", rec.Body.String())
+
+	rec = f.do(t, http.MethodDelete, "/agent-network/settings", "")
+	require.Equal(t, http.StatusOK, rec.Code,
+		"delete with both guards clear must succeed: got %d body=%s", rec.Code, rec.Body.String())
+
+	rec = f.do(t, http.MethodGet, "/agent-network/settings", "")
+	require.Equal(t, http.StatusOK, rec.Code)
+	var after api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &after))
+	assert.Empty(t, after.Endpoint, "a deleted account must read as unbootstrapped defaults")
+	assert.False(t, after.EnablePromptCollection, "the deleted row's toggles must not linger")
+
+	rec = f.do(t, http.MethodPost, "/agent-network/settings", `{"proxy_address": "eu.proxy.netbird.io"}`)
+	require.Equal(t, http.StatusOK, rec.Code, "re-bootstrap after delete must succeed: %s", rec.Body.String())
+	var second api.AgentNetworkSettings
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &second))
+	require.NotEmpty(t, second.Endpoint, "the fresh bootstrap must allocate an endpoint")
+	assert.True(t, strings.HasSuffix(second.Endpoint, ".eu.proxy.netbird.io"),
+		"the fresh endpoint must hang beneath the requested proxy address: %s", second.Endpoint)
+	assert.False(t, second.EnablePromptCollection,
+		"the fresh row must carry bootstrap defaults, not the deleted row's toggles")
+	assert.NotNil(t, second.CreatedAt, "the fresh row is persisted and carries timestamps")
 }

management/internals/modules/agentnetwork/labelgen/adjectives.go

@@ -0,0 +1,37 @@
+package labelgen
+
+// adjectives is the descriptor half of a generated label. It pairs with the
+// noun pool in words.go to form `<adjective>-<noun>` labels, and is kept
+// separate because words.go is almost entirely nouns — drawing both halves
+// from it produced unreadable pairs like "millet-hammock". Entries are
+// lowercase ASCII, 4-12 chars, free of hyphens and digits, screened for
+// offensive/brand/region-specific terms, and disjoint from the noun pool
+// (enforced by TestAdjectives_AreDisjointFromNouns).
+var adjectives = []string{
+	"able", "active", "adept", "agile", "airy", "alert", "amiable", "ample",
+	"ancient", "ardent", "artful", "astute", "balmy", "blithe", "bold", "bonny",
+	"brave", "breezy", "brisk", "bubbly", "buoyant", "bushy", "candid", "canny",
+	"cheery", "chilly", "chipper", "chunky", "civil", "classic", "clever", "comely",
+	"compact", "cordial", "cosmic", "courtly", "crafty", "creamy", "crisp", "cuddly",
+	"curious", "dainty", "dapper", "daring", "dashing", "deft", "dewy", "diligent",
+	"downy", "dreamy", "dulcet", "durable", "dusky", "eager", "earnest", "earthy",
+	"easy", "elated", "elegant", "epic", "fabled", "faithful", "fancy", "fearless",
+	"feisty", "fervent", "fleet", "fluffy", "fond", "frisky", "frosty", "gallant",
+	"genial", "genteel", "gentle", "giddy", "gilded", "glad", "glassy", "gleaming",
+	"glossy", "graceful", "grand", "grainy", "hale", "hardy", "hearty", "hefty",
+	"honest", "hopeful", "humble", "hushed", "immense", "jaunty", "jolly", "jovial",
+	"joyful", "jubilant", "keen", "kindly", "kindred", "lanky", "leafy", "limber",
+	"lively", "lofty", "loyal", "lucent", "lucid", "luminous", "lush", "maroon",
+	"mellow", "merry", "mighty", "mindful", "mirthful", "misty", "modest", "muted",
+	"nifty", "nimble", "noble", "patient", "peaceful", "pearly", "peppy", "perky",
+	"petite", "placid", "playful", "pleasant", "plucky", "plush", "polite", "posh",
+	"prancing", "pristine", "prompt", "proud", "prudent", "quaint", "quick", "quirky",
+	"radiant", "ready", "regal", "restful", "robust", "rosy", "ruddy", "rugged",
+	"sandy", "satin", "saucy", "savvy", "sedate", "serene", "shady", "shiny",
+	"silken", "silky", "sincere", "sleek", "slender", "smart", "smooth", "snappy",
+	"snug", "soaring", "sparkly", "spiffy", "spirited", "sprightly", "spry", "stalwart",
+	"stately", "steady", "sterling", "stoic", "stormy", "stout", "sturdy", "sunlit",
+	"supple", "svelte", "tawny", "tender", "tidy", "timeless", "trusty", "upbeat",
+	"urbane", "valiant", "vast", "vernal", "vibrant", "vintage", "whimsy", "willing",
+	"windy", "winsome", "wintry", "witty", "worthy", "zesty", "zippy",
+}

management/internals/modules/agentnetwork/labelgen/labelgen.go

@@ -64,3 +64,20 @@ func PickUnique(rng *rand.Rand, taken map[string]struct{}, fallbackSuffix string
 	w := pool[rng.Intn(len(pool))]
 	return fmt.Sprintf("%s-%s", w, fallbackSuffix)
 }
+
+// PickTuple returns an adjective-noun label such as "brave-otter". It is still
+// a single DNS label.
+//
+// Unlike PickUnique it takes no `taken` set and has no fallback suffix. The
+// noun pool holds 857 entries, which is ample per cluster but a hard ceiling
+// once labels must be unique across one shared zone; pairing an adjective with
+// a noun spans len(adjectives) * 857 instead. Uniqueness is enforced by a
+// database constraint and retried by the caller, rather than guessed from a
+// pre-read set that a concurrent allocation can invalidate.
+func PickTuple(rng *rand.Rand) string {
+	nouns := uniqueWords()
+	if len(nouns) == 0 || len(adjectives) == 0 {
+		return ""
+	}
+	return adjectives[rng.Intn(len(adjectives))] + "-" + nouns[rng.Intn(len(nouns))]
+}

management/internals/modules/agentnetwork/labelgen/labelgen_test.go

@@ -99,3 +99,82 @@ func TestUniqueWords_DropsDuplicates(t *testing.T) {
 	}
 	assert.GreaterOrEqual(t, len(pool), 500, "Pool must contain at least 500 unique words")
 }
+
+// TestPickTuple_ShapeAndPoolMembership locks the wire-visible shape: an
+// adjective and a noun, each from its own pool, joined by a single hyphen so
+// the result stays one DNS label.
+func TestPickTuple_ShapeAndPoolMembership(t *testing.T) {
+	nouns := uniqueWords()
+	inNouns := make(map[string]struct{}, len(nouns))
+	for _, w := range nouns {
+		inNouns[w] = struct{}{}
+	}
+	inAdjectives := make(map[string]struct{}, len(adjectives))
+	for _, a := range adjectives {
+		inAdjectives[a] = struct{}{}
+	}
+
+	rng := rand.New(rand.NewSource(7))
+	for i := 0; i < 200; i++ {
+		got := PickTuple(rng)
+
+		parts := strings.Split(got, "-")
+		require.Len(t, parts, 2, "PickTuple must produce exactly two hyphen-joined words; got %q", got)
+
+		_, adjOK := inAdjectives[parts[0]]
+		assert.True(t, adjOK, "First half must be an adjective; %q not in adjectives (from %q)", parts[0], got)
+		_, nounOK := inNouns[parts[1]]
+		assert.True(t, nounOK, "Second half must be a noun; %q not in words (from %q)", parts[1], got)
+
+		assert.LessOrEqual(t, len(got), 63, "Label must fit a DNS label; got %q (%d chars)", got, len(got))
+	}
+}
+
+// TestAdjectives_AreDisjointFromNouns keeps the namespace a clean product and
+// prevents nonsense like "azure-azure": a handful of the noun pool's entries
+// are adjectival, and any overlap would let the same word land on both sides.
+func TestAdjectives_AreDisjointFromNouns(t *testing.T) {
+	nouns := make(map[string]struct{}, len(uniqueWords()))
+	for _, w := range uniqueWords() {
+		nouns[w] = struct{}{}
+	}
+	for _, a := range adjectives {
+		_, clash := nouns[a]
+		assert.False(t, clash, "Adjective %q also appears in the noun pool; remove it from one list", a)
+	}
+}
+
+// TestAdjectives_AreDNSSafeAndDeduplicated mirrors the curation contract stated
+// in words.go: lowercase ASCII, 4-12 chars, no digits or hyphens, no repeats.
+func TestAdjectives_AreDNSSafeAndDeduplicated(t *testing.T) {
+	seen := make(map[string]struct{}, len(adjectives))
+	for _, a := range adjectives {
+		_, dup := seen[a]
+		assert.False(t, dup, "Duplicate adjective %q", a)
+		seen[a] = struct{}{}
+
+		assert.Regexp(t, `^[a-z]{4,12}$`, a, "Adjective %q must be 4-12 lowercase ASCII letters", a)
+	}
+	assert.Greater(t, len(adjectives), 150, "Adjective pool too small to give a useful namespace")
+}
+
+// TestPickTuple_DeterministicWithSeededRng documents that generation is a pure
+// function of the rng, which is what makes allocation retries reproducible in tests.
+func TestPickTuple_DeterministicWithSeededRng(t *testing.T) {
+	a := PickTuple(rand.New(rand.NewSource(42)))
+	b := PickTuple(rand.New(rand.NewSource(42)))
+	assert.Equal(t, a, b, "Same seed must yield the same tuple")
+}
+
+// TestPickTuple_SpansALargeNamespace guards the reason we moved to tuples: a
+// single-word pool caps the GLOBAL namespace at 857. Drawing many tuples must
+// yield overwhelmingly distinct values.
+func TestPickTuple_SpansALargeNamespace(t *testing.T) {
+	rng := rand.New(rand.NewSource(11))
+	seen := make(map[string]struct{}, 2000)
+	for i := 0; i < 2000; i++ {
+		seen[PickTuple(rng)] = struct{}{}
+	}
+	assert.Greater(t, len(seen), 1900,
+		"2000 draws should be nearly all distinct across a ~200k namespace; got %d unique", len(seen))
+}

management/internals/modules/agentnetwork/manager.go

@@ -13,6 +13,7 @@ import (
 	log "github.com/sirupsen/logrus"
 
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/labelgen"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/modeldiscovery"
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/sessionkey"
@@ -22,7 +23,6 @@ import (
 	"github.com/netbirdio/netbird/management/server/permissions/modules"
 	"github.com/netbirdio/netbird/management/server/permissions/operations"
 	"github.com/netbirdio/netbird/management/server/store"
-	"github.com/netbirdio/netbird/shared/management/proto"
 	"github.com/netbirdio/netbird/shared/management/status"
 )
 
@@ -48,9 +48,10 @@ func ensureSessionKeys(p *types.Provider) error {
 type Manager interface {
 	GetAllProviders(ctx context.Context, accountID, userID string) ([]*types.Provider, error)
 	GetProvider(ctx context.Context, accountID, userID, providerID string) (*types.Provider, error)
-	CreateProvider(ctx context.Context, userID string, provider *types.Provider, bootstrapCluster string) (*types.Provider, error)
+	CreateProvider(ctx context.Context, userID string, provider *types.Provider) (*types.Provider, error)
 	UpdateProvider(ctx context.Context, userID string, provider *types.Provider) (*types.Provider, error)
 	DeleteProvider(ctx context.Context, accountID, userID, providerID string) error
+	DiscoverProviderModels(ctx context.Context, accountID, userID string, req modeldiscovery.Request, recordID string) ([]modeldiscovery.Model, error)
 
 	GetAllPolicies(ctx context.Context, accountID, userID string) ([]*types.Policy, error)
 	GetPolicy(ctx context.Context, accountID, userID, policyID string) (*types.Policy, error)
@@ -71,7 +72,9 @@ type Manager interface {
 	DeleteBudgetRule(ctx context.Context, accountID, userID, ruleID string) error
 
 	GetSettings(ctx context.Context, accountID, userID string) (*types.Settings, error)
+	CreateSettings(ctx context.Context, userID string, settings *types.Settings, proxyAddress, endpoint string) (*types.Settings, error)
 	UpdateSettings(ctx context.Context, userID string, settings *types.Settings) (*types.Settings, error)
+	DeleteSettings(ctx context.Context, accountID, userID string) error
 
 	ListConsumption(ctx context.Context, accountID, userID string) ([]*types.Consumption, error)
 	ListAccessLogs(ctx context.Context, accountID, userID string, filter types.AgentNetworkAccessLogFilter) ([]*types.AgentNetworkAccessLog, int64, error)
@@ -122,12 +125,20 @@ type managerImpl struct {
 	permissionsManager permissions.Manager
 	proxyController    proxy.Controller
 
+	// modelDiscovery queries vendors for the models a credential can reach.
+	// A field rather than a package call so tests can drive it without
+	// reaching the network.
+	//
+	// One instance serves every request for the process's lifetime, so its
+	// fields must stay read-only after construction: lazy initialisation
+	// inside Fetch or httpClient would race across request goroutines.
+	modelDiscovery *modeldiscovery.Client
+
 	// reconcileCache holds the last set of synthesised proxy mappings
-	// per account so reconcile can emit precise Create/Update/Delete
-	// updates instead of a full re-push on every mutation. Keyed by
-	// accountID, then by synthesised service ID.
+	// per account, each paired with the proxy that served it, so a change
+	// of serving proxy can be diffed without re-deriving it.
 	reconcileMu    sync.Mutex
-	reconcileCache map[string]map[string]*proto.ProxyMapping
+	reconcileCache map[string]map[string]syntheticMapping
 
 	// labelRngMu guards labelRng. PickUnique consumes math/rand.Source
 	// state; concurrent provider creates would otherwise race.
@@ -151,7 +162,8 @@ func NewManager(
 		accountManager:     accountManager,
 		permissionsManager: permissionsManager,
 		proxyController:    proxyController,
-		reconcileCache:     make(map[string]map[string]*proto.ProxyMapping),
+		modelDiscovery:     &modeldiscovery.Client{},
+		reconcileCache:     make(map[string]map[string]syntheticMapping),
 		labelRng:           rand.New(rand.NewSource(time.Now().UnixNano())),
 	}
 }
@@ -170,18 +182,45 @@ func (m *managerImpl) GetProvider(ctx context.Context, accountID, userID, provid
 	return m.store.GetAgentNetworkProviderByID(ctx, store.LockingStrengthNone, accountID, providerID)
 }
 
-// CreateProvider persists a new provider for the account. bootstrapCluster
-// is used only when the per-account agent-network Settings row hasn't
-// been created yet; otherwise it is ignored (the cluster is pinned on
-// Settings and every provider in the account routes through it).
-func (m *managerImpl) CreateProvider(ctx context.Context, userID string, provider *types.Provider, bootstrapCluster string) (*types.Provider, error) {
-	if err := m.requirePermission(ctx, provider.AccountID, userID, modules.AgentNetworkProviders, operations.Create); err != nil {
+// DiscoverProviderModels asks the vendor which models a credential can reach.
+//
+// recordID, when set, names an existing provider whose stored credential and
+// upstream are used instead of the ones in req — so the dashboard can refresh
+// the list without ever holding the key.
+//
+// Gated on Create rather than Read: this spends the operator's credential
+// against a third party, which is not something a read-only role should be
+// able to make the server do. That one check also covers reading the stored
+// record — Create is strictly stronger than Read here, and the lookup is
+// scoped to accountID, so another account's record is never reachable.
+func (m *managerImpl) DiscoverProviderModels(ctx context.Context, accountID, userID string, req modeldiscovery.Request, recordID string) ([]modeldiscovery.Model, error) {
+	if err := m.requirePermission(ctx, accountID, userID, modules.AgentNetworkProviders, operations.Create); err != nil {
 		return nil, err
 	}
-	if strings.TrimSpace(bootstrapCluster) != "" {
-		if err := m.requireSettingsBootstrapPermission(ctx, provider.AccountID, userID); err != nil {
+
+	if recordID != "" {
+		record, err := m.store.GetAgentNetworkProviderByID(ctx, store.LockingStrengthNone, accountID, recordID)
+		if err != nil {
 			return nil, err
 		}
+		// The catalog id comes from the stored record too: letting the caller
+		// name a different one would run a provider's credential against
+		// whichever vendor endpoint they picked.
+		req.CatalogID = record.ProviderID
+		req.UpstreamURL = record.UpstreamURL
+		req.APIKey = record.APIKey
+	}
+
+	return m.modelDiscovery.Fetch(ctx, req)
+}
+
+// CreateProvider persists a new provider for the account. Providers have no
+// settings side effects: the account's endpoint is bootstrapped separately and
+// explicitly via CreateSettings, and every provider in the account routes
+// through it.
+func (m *managerImpl) CreateProvider(ctx context.Context, userID string, provider *types.Provider) (*types.Provider, error) {
+	if err := m.requirePermission(ctx, provider.AccountID, userID, modules.AgentNetworkProviders, operations.Create); err != nil {
+		return nil, err
 	}
 
 	// An empty api_key would silently produce a synthesised service
@@ -206,16 +245,6 @@ func (m *managerImpl) CreateProvider(ctx context.Context, userID string, provide
 		return nil, fmt.Errorf("save agent network provider: %w", err)
 	}
 
-	if strings.TrimSpace(bootstrapCluster) != "" {
-		if _, err := m.bootstrapSettingsIfNeeded(ctx, m.store, provider.AccountID, bootstrapCluster); err != nil {
-			// The provider create has already succeeded; logging the
-			// bootstrap miss matches the plan's PoC behaviour. The synth
-			// path treats a missing settings row as a no-op, and the next
-			// provider create retries the bootstrap.
-			log.WithContext(ctx).Debugf("agent-network bootstrap settings for account %s on cluster %s: %v", provider.AccountID, bootstrapCluster, err)
-		}
-	}
-
 	m.accountManager.StoreEvent(ctx, userID, provider.ID, provider.AccountID, activity.AgentNetworkProviderCreated, provider.EventMeta())
 	m.reconcile(ctx, provider.AccountID)
 
@@ -560,52 +589,44 @@ func (m *managerImpl) DeleteBudgetRule(ctx context.Context, accountID, userID, r
 }
 
 // UpdateSettings replaces the mutable account-level settings — the collection
-// toggles and retention — on the account's row. When the account has no
-// settings row yet, a non-empty settings.Cluster bootstraps one (same path as
-// first provider create); without it the update fails with NotFound. On an
-// existing row the cluster and subdomain are immutable: a differing
-// settings.Cluster is rejected rather than silently ignored so callers never
-// observe a value other than what they sent. Because the collection toggles
-// change the synthesised service config (prompt-capture gating, access-log
-// emission), a reconcile is triggered so the proxy and peer network maps
-// converge on the new state.
+// toggles and retention — on the account's row. The identity fields (Domain,
+// ProxyAddress) are assigned at bootstrap (CreateSettings) and immutable: the
+// request carries them, matching the PUT convention of every other endpoint,
+// but they are only compared against the stored row — a request carrying
+// different values is rejected, and the stored values are never overwritten.
+// When the account has no settings row yet the update fails with NotFound.
+// Because the collection toggles change the synthesised service config
+// (prompt-capture gating, access-log emission), a reconcile is triggered so
+// the proxy and peer network maps converge on the new state.
 func (m *managerImpl) UpdateSettings(ctx context.Context, userID string, settings *types.Settings) (*types.Settings, error) {
 	if err := m.requirePermission(ctx, settings.AccountID, userID, modules.AgentNetworkSettings, operations.Update); err != nil {
 		return nil, err
 	}
 
-	requestedCluster := strings.TrimSpace(settings.Cluster)
-
 	// The row lock from LockingStrengthUpdate only holds for the duration of
-	// the surrounding transaction, so the read, the cluster-immutability
-	// check, and the save must share one — otherwise concurrent PUTs could
-	// interleave between them.
+	// the surrounding transaction, so the read and the save must share one —
+	// otherwise concurrent PUTs could interleave between them.
 	var updated *types.Settings
 	err := m.store.ExecuteInTransaction(ctx, func(tx store.Store) error {
 		existing, err := tx.GetAgentNetworkSettings(ctx, store.LockingStrengthUpdate, settings.AccountID)
 		switch {
 		case err == nil:
-			if requestedCluster != "" && requestedCluster != existing.Cluster {
-				return status.Errorf(status.InvalidArgument, "cluster is immutable once assigned (current: %s)", existing.Cluster)
-			}
 		case isNotFound(err):
-			if requestedCluster == "" {
-				return status.Errorf(status.NotFound, "agent network settings have not been bootstrapped yet; pass cluster to bootstrap them, or create a provider with bootstrap_cluster set")
-			}
-			// Bootstrapping pins the cluster and subdomain — a settings
-			// create on top of the update the caller already passed, matching
-			// the gate on the provider-create bootstrap path.
-			if err := m.requirePermission(ctx, settings.AccountID, userID, modules.AgentNetworkSettings, operations.Create); err != nil {
-				return err
-			}
-			existing, err = m.bootstrapSettingsIfNeeded(ctx, tx, settings.AccountID, requestedCluster)
-			if err != nil {
-				return err
-			}
+			return status.Errorf(status.NotFound, "agent network settings have not been bootstrapped yet; POST /api/agent-network/settings to bootstrap them")
 		default:
 			return fmt.Errorf("get agent network settings: %w", err)
 		}
 
+		// The identity echo is compared leniently (trimmed, case-insensitive):
+		// the stored values are normalized lowercase, and a client replaying a
+		// GET response must never be rejected over casing it didn't choose.
+		if !hostnamesEquivalent(settings.Domain, existing.Domain) {
+			return status.Errorf(status.InvalidArgument, "endpoint is immutable: it must match the assigned endpoint %q; delete the settings to release it and bootstrap again", existing.Domain)
+		}
+		if !hostnamesEquivalent(settings.ProxyAddress, existing.ProxyAddress) {
+			return status.Errorf(status.InvalidArgument, "proxy_address is immutable: it must match the assigned proxy address %q; delete the settings to release it and bootstrap again", existing.ProxyAddress)
+		}
+
 		existing.EnableLogCollection = settings.EnableLogCollection
 		existing.EnablePromptCollection = settings.EnablePromptCollection
 		existing.RedactPii = settings.RedactPii
@@ -632,6 +653,83 @@ func (m *managerImpl) UpdateSettings(ctx context.Context, userID string, setting
 	return updated, nil
 }
 
+// hostnamesEquivalent reports whether a caller-supplied hostname names the
+// same host as a stored (normalized, lowercase) one: equal after trimming and
+// case folding. No structural validation — an arbitrary mismatch and a
+// malformed value are both simply "not the assigned value".
+func hostnamesEquivalent(supplied, stored string) bool {
+	return strings.EqualFold(strings.TrimSpace(supplied), stored)
+}
+
+// DeleteSettings removes the account's settings row, releasing the endpoint.
+// Two guards make this a bootstrap-repair operation rather than a way to tear
+// down a serving gateway, both re-checked under the row lock:
+//
+//   - No Agent Network providers may exist for the account. Providers route
+//     through the endpoint; delete them first.
+//   - No proxy may be actively serving the endpoint — that is, no active proxy
+//     declares the endpoint hostname as its cluster address. This is the
+//     dedicated (self-addressed) shape's guard: the proxy at the address IS
+//     this account's gateway. A labeled endpoint hangs beneath a shared
+//     cluster's address, and with the account's providers already gone the
+//     shared proxy serves nothing of the account's, so the parent cluster
+//     being up does not block the delete.
+//
+// Bootstrapping again after a delete allocates fresh — the released hostname
+// is not reserved. That full-reset semantic is what gives clients that model
+// immutability as replace-on-change (e.g. Terraform's RequiresReplace) a real
+// path: tear down providers, delete, re-create.
+func (m *managerImpl) DeleteSettings(ctx context.Context, accountID, userID string) error {
+	if err := m.requirePermission(ctx, accountID, userID, modules.AgentNetworkSettings, operations.Delete); err != nil {
+		return err
+	}
+
+	var deleted *types.Settings
+	err := m.store.ExecuteInTransaction(ctx, func(tx store.Store) error {
+		existing, err := tx.GetAgentNetworkSettings(ctx, store.LockingStrengthUpdate, accountID)
+		switch {
+		case err == nil:
+		case isNotFound(err):
+			return status.Errorf(status.NotFound, "agent network settings have not been bootstrapped yet; there is nothing to delete")
+		default:
+			return fmt.Errorf("get agent network settings: %w", err)
+		}
+
+		providers, err := tx.GetAccountAgentNetworkProviders(ctx, store.LockingStrengthNone, accountID)
+		if err != nil {
+			return fmt.Errorf("get agent network providers: %w", err)
+		}
+		if len(providers) > 0 {
+			return status.Errorf(status.PreconditionFailed, "agent network settings cannot be deleted while %d provider(s) exist; delete the providers first", len(providers))
+		}
+
+		serving, err := tx.HasActiveProxyAtClusterAddress(ctx, existing.Domain)
+		if err != nil {
+			return fmt.Errorf("check for a proxy serving the endpoint: %w", err)
+		}
+		if serving {
+			return status.Errorf(status.PreconditionFailed, "agent network settings cannot be deleted while a proxy is actively serving the endpoint %q", existing.Domain)
+		}
+
+		if err := tx.DeleteAgentNetworkSettings(ctx, accountID); err != nil {
+			return fmt.Errorf("delete agent network settings: %w", err)
+		}
+		deleted = existing
+		return nil
+	})
+	if err != nil {
+		return err
+	}
+
+	m.accountManager.StoreEvent(ctx, userID, accountID, accountID, activity.AgentNetworkSettingsDeleted, map[string]any{
+		"endpoint":      deleted.Domain,
+		"proxy_address": deleted.ProxyAddress,
+	})
+	m.reconcile(ctx, accountID)
+
+	return nil
+}
+
 // isNotFound reports whether err is a status.NotFound error.
 func isNotFound(err error) bool {
 	var sErr *status.Error
@@ -678,74 +776,162 @@ func (m *managerImpl) GetSettings(ctx context.Context, accountID, userID string)
 	}
 }
 
-// requireSettingsBootstrapPermission gates the one-time settings bootstrap a
-// first provider create performs. Pinning the account's cluster and subdomain
-// is a settings write, so it needs the settings permission on top of the
-// provider one. No-op once the settings row exists.
-func (m *managerImpl) requireSettingsBootstrapPermission(ctx context.Context, accountID, userID string) error {
-	_, err := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, accountID)
-	if err == nil {
-		return nil
-	}
-	if !isNotFound(err) {
-		return fmt.Errorf("get agent network settings: %w", err)
-	}
-	return m.requirePermission(ctx, accountID, userID, modules.AgentNetworkSettings, operations.Create)
-}
+// maxDomainAllocationAttempts bounds the label search when bootstrapping a
+// labeled endpoint. Package-level (rather than function-local) so tests can
+// assert on the exhaustion path without duplicating the literal.
+const maxDomainAllocationAttempts = 10
 
-// bootstrapSettingsIfNeeded creates the per-account agent-network
-// settings row when missing. The cluster comes from the create-time
-// hint the dashboard sends (auto-picked from the active cluster list);
-// the subdomain is picked from the curated wordlist avoiding
-// collisions on the same cluster. Idempotent: if a row already exists
-// it is returned untouched and the hint is ignored. st is the store to
-// operate on — pass the transaction store when calling from within one.
-func (m *managerImpl) bootstrapSettingsIfNeeded(ctx context.Context, st store.Store, accountID, providerCluster string) (*types.Settings, error) {
-	if accountID == "" {
-		return nil, fmt.Errorf("bootstrap settings: account id is required")
+// CreateSettings bootstraps the per-account settings row, assigning the
+// account's immutable endpoint. Exactly one of proxyAddress and endpoint must
+// be non-empty: proxyAddress allocates a labeled endpoint one label beneath
+// the given cluster address; endpoint claims the given hostname verbatim as a
+// self-addressed (dedicated) endpoint — a legitimate claim before any proxy
+// declares the address (address-first). settings carries the account ID and
+// the initial collection toggles; its identity fields are assigned here.
+func (m *managerImpl) CreateSettings(ctx context.Context, userID string, settings *types.Settings, proxyAddress, endpoint string) (*types.Settings, error) {
+	if settings == nil || settings.AccountID == "" {
+		return nil, status.Errorf(status.InvalidArgument, "account id is required")
 	}
-	if strings.TrimSpace(providerCluster) == "" {
-		return nil, fmt.Errorf("bootstrap settings: provider cluster is required")
+	if err := m.requirePermission(ctx, settings.AccountID, userID, modules.AgentNetworkSettings, operations.Create); err != nil {
+		return nil, err
 	}
 
-	existing, err := st.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, accountID)
-	if err == nil {
-		return existing, nil
+	hasProxyAddress := strings.TrimSpace(proxyAddress) != ""
+	hasEndpoint := strings.TrimSpace(endpoint) != ""
+	if hasProxyAddress == hasEndpoint {
+		return nil, status.Errorf(status.InvalidArgument, "exactly one of proxy_address and endpoint is required")
 	}
-	if !isNotFound(err) {
+
+	// Fail fast on an existing row for a clean 409; the insert below stays
+	// the authority against concurrent bootstraps (the primary key wins).
+	if _, err := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, settings.AccountID); err == nil {
+		return nil, status.Errorf(status.AlreadyExists, "agent network settings already bootstrapped for account %s", settings.AccountID)
+	} else if !isNotFound(err) {
 		return nil, fmt.Errorf("get agent network settings: %w", err)
 	}
 
-	siblings, err := st.GetAgentNetworkSettingsByCluster(ctx, store.LockingStrengthNone, providerCluster)
-	if err != nil {
-		return nil, fmt.Errorf("list agent network settings on cluster: %w", err)
-	}
-	taken := make(map[string]struct{}, len(siblings))
-	for _, s := range siblings {
-		taken[s.Subdomain] = struct{}{}
-	}
-
-	suffix := accountID
-	if len(suffix) > 4 {
-		suffix = suffix[:4]
-	}
-
-	m.labelRngMu.Lock()
-	subdomain := labelgen.PickUnique(m.labelRng, taken, suffix)
-	m.labelRngMu.Unlock()
-
 	now := time.Now().UTC()
-	settings := types.DefaultSettings(accountID)
-	settings.Cluster = providerCluster
-	settings.Subdomain = subdomain
 	settings.CreatedAt = now
 	settings.UpdatedAt = now
-	if err := st.SaveAgentNetworkSettings(ctx, settings); err != nil {
-		return nil, fmt.Errorf("save agent network settings: %w", err)
+
+	var err error
+	if hasEndpoint {
+		err = m.bootstrapSelfAddressed(ctx, settings, endpoint)
+	} else {
+		err = m.bootstrapLabeled(ctx, settings, proxyAddress)
 	}
+	if err != nil {
+		return nil, err
+	}
+
+	m.accountManager.StoreEvent(ctx, userID, settings.AccountID, settings.AccountID, activity.AgentNetworkSettingsUpdated, map[string]any{
+		"bootstrapped": true,
+		"endpoint":     settings.Domain,
+		"dedicated":    settings.Dedicated(),
+	})
+	m.reconcile(ctx, settings.AccountID)
+
 	return settings, nil
 }
 
+// bootstrapSelfAddressed claims the given hostname as the account's endpoint,
+// served only by a proxy declaring exactly that address (Domain ==
+// ProxyAddress). The domain unique index is the arbiter of availability.
+func (m *managerImpl) bootstrapSelfAddressed(ctx context.Context, settings *types.Settings, endpoint string) error {
+	hostname, err := types.NormalizeHostname(endpoint)
+	if err != nil {
+		return status.Errorf(status.InvalidArgument, "invalid endpoint: %s", err)
+	}
+
+	settings.Domain = hostname
+	settings.ProxyAddress = hostname
+	if err := m.store.CreateAgentNetworkSettings(ctx, settings); err != nil {
+		if isUniqueConstraintError(err) {
+			// The violation is either the account primary key (a concurrent
+			// bootstrap for the same account won) or the domain index
+			// (another account holds the hostname). Distinguish by re-read.
+			if _, getErr := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, settings.AccountID); getErr == nil {
+				return status.Errorf(status.AlreadyExists, "agent network settings already bootstrapped for account %s", settings.AccountID)
+			}
+			return status.Errorf(status.AlreadyExists, "endpoint %s is already taken", hostname)
+		}
+		return fmt.Errorf("create agent network settings: %w", err)
+	}
+	return nil
+}
+
+// bootstrapLabeled allocates a labeled endpoint one label beneath the given
+// cluster address: Domain = <label>.<proxyAddress>, served by whichever proxy
+// declares the parent. Labels are adjective-noun tuples; a candidate is
+// checked by read and the domain unique index stays the authority, so a
+// concurrent allocation of the same tuple surfaces as a unique violation and
+// another tuple is drawn.
+func (m *managerImpl) bootstrapLabeled(ctx context.Context, settings *types.Settings, proxyAddress string) error {
+	parent, err := types.NormalizeHostname(proxyAddress)
+	if err != nil {
+		return status.Errorf(status.InvalidArgument, "invalid proxy_address: %s", err)
+	}
+
+	for attempt := 1; attempt <= maxDomainAllocationAttempts; attempt++ {
+		m.labelRngMu.Lock()
+		label := labelgen.PickTuple(m.labelRng)
+		m.labelRngMu.Unlock()
+		if label == "" {
+			// Only reachable if either word pool were emptied. An empty label
+			// would produce a broken endpoint like ".example.com", so fail
+			// loudly rather than looping or inserting.
+			return fmt.Errorf("allocate agent network endpoint for account %s: label generator returned an empty label", settings.AccountID)
+		}
+
+		candidate, err := types.NormalizeHostname(label + "." + parent)
+		if err != nil {
+			return status.Errorf(status.InvalidArgument, "proxy_address leaves no room for a label: %s", err)
+		}
+
+		_, err = m.store.GetAgentNetworkSettingsByDomain(ctx, store.LockingStrengthNone, candidate)
+		if err == nil {
+			log.WithContext(ctx).Tracef("agent-network endpoint %q taken, retrying (attempt %d/%d)", candidate, attempt, maxDomainAllocationAttempts)
+			continue
+		}
+		if !isNotFound(err) {
+			return fmt.Errorf("check agent network endpoint availability: %w", err)
+		}
+
+		settings.Domain = candidate
+		settings.ProxyAddress = parent
+		if err := m.store.CreateAgentNetworkSettings(ctx, settings); err != nil {
+			if isUniqueConstraintError(err) {
+				// A concurrent bootstrap for the same account may have won on
+				// the primary key — return the conflict. A lost race on the
+				// domain index just means the tuple was taken between the
+				// read and the insert: draw another.
+				if _, getErr := m.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, settings.AccountID); getErr == nil {
+					return status.Errorf(status.AlreadyExists, "agent network settings already bootstrapped for account %s", settings.AccountID)
+				}
+				log.WithContext(ctx).Tracef("agent-network endpoint %q lost an allocation race, retrying (attempt %d/%d)", candidate, attempt, maxDomainAllocationAttempts)
+				continue
+			}
+			return fmt.Errorf("create agent network settings: %w", err)
+		}
+		return nil
+	}
+
+	return fmt.Errorf("allocate agent network endpoint for account %s: %d attempts exhausted", settings.AccountID, maxDomainAllocationAttempts)
+}
+
+// isUniqueConstraintError reports whether err is a database unique-constraint
+// violation, matched on the driver message because CreateAgentNetworkSettings
+// deliberately returns the driver error unwrapped.
+func isUniqueConstraintError(err error) bool {
+	if err == nil {
+		return false
+	}
+	msg := err.Error()
+	return strings.Contains(msg, "(SQLSTATE 23505)") || // postgres
+		strings.Contains(msg, "Error 1062 (23000)") || // mysql
+		strings.Contains(msg, "UNIQUE constraint failed") // sqlite
+}
+
 // ListConsumption returns every consumption row recorded for the
 // account, ordered window-newest-first. Backs the dashboard's basic
 // counter view; permission gate is the same Read role that gates
@@ -875,11 +1061,15 @@ func (*mockManager) GetAllProviders(_ context.Context, _, _ string) ([]*types.Pr
 	return []*types.Provider{}, nil
 }
 
+func (*mockManager) DiscoverProviderModels(_ context.Context, _, _ string, _ modeldiscovery.Request, _ string) ([]modeldiscovery.Model, error) {
+	return nil, nil
+}
+
 func (*mockManager) GetProvider(_ context.Context, _, _, _ string) (*types.Provider, error) {
 	return &types.Provider{}, nil
 }
 
-func (*mockManager) CreateProvider(_ context.Context, _ string, p *types.Provider, _ string) (*types.Provider, error) {
+func (*mockManager) CreateProvider(_ context.Context, _ string, p *types.Provider) (*types.Provider, error) {
 	return p, nil
 }
 
@@ -947,10 +1137,23 @@ func (*mockManager) GetSettings(_ context.Context, accountID, _ string) (*types.
 	return types.DefaultSettings(accountID), nil
 }
 
+func (*mockManager) CreateSettings(_ context.Context, _ string, s *types.Settings, proxyAddress, endpoint string) (*types.Settings, error) {
+	if endpoint != "" {
+		s.Domain = endpoint
+		s.ProxyAddress = endpoint
+	} else {
+		s.Domain = "mock." + proxyAddress
+		s.ProxyAddress = proxyAddress
+	}
+	return s, nil
+}
+
 func (*mockManager) UpdateSettings(_ context.Context, _ string, s *types.Settings) (*types.Settings, error) {
 	return s, nil
 }
 
+func (*mockManager) DeleteSettings(_ context.Context, _, _ string) error { return nil }
+
 func (*mockManager) ListConsumption(_ context.Context, _, _ string) ([]*types.Consumption, error) {
 	return nil, nil
 }

management/internals/modules/agentnetwork/modeldiscovery/discovery.go

@@ -0,0 +1,469 @@
+// Package modeldiscovery asks a vendor which models an operator's own
+// credential can reach, so the provider form can offer a live list instead of
+// only the catalog's hand-curated one.
+//
+// The catalog cannot know two things that matter. It goes stale — its entries
+// carry comments tracking which models a vendor retired on which date — and it
+// cannot see an account: which OpenAI models an org is entitled to, which
+// Bedrock inference profiles a given account and region hold, which Vertex
+// models a project has enabled. Those are exactly the facts an operator needs
+// when filling in a provider record, and only the vendor has them.
+//
+// The vendor is authoritative for the model ID. The catalog remains
+// authoritative for pricing, and a discovered model the catalog cannot price
+// is reported as such rather than silently registered at a rate of zero.
+package modeldiscovery
+
+import (
+	"context"
+	"encoding/base64"
+	"errors"
+	"fmt"
+	"io"
+	"net"
+	"net/http"
+	"net/netip"
+	"net/url"
+	"strings"
+	"syscall"
+	"time"
+
+	"golang.org/x/oauth2/google"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/catalog"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/pricing"
+)
+
+const (
+	// fetchTimeout bounds one vendor call end to end. A listing is a single
+	// small GET; anything slower is a vendor problem and the operator is
+	// waiting on a form.
+	fetchTimeout = 8 * time.Second
+	// maxListingBytes bounds the response we will buffer. The largest real
+	// listing observed is Bedrock's foundation-model catalogue at ~70KB, so
+	// this is a wide margin over anything legitimate.
+	maxListingBytes = 2 << 20
+	// gcpScope matches the scope llm_router mints Vertex tokens under, so a
+	// credential that works for discovery works for inference too.
+	gcpScope = "https://www.googleapis.com/auth/cloud-platform"
+	// vertexKeyfilePrefix marks an api_key that is a base64 service-account
+	// JSON key rather than a bearer token.
+	vertexKeyfilePrefix = "keyfile::"
+)
+
+// ErrNoDiscovery is returned for a catalog entry that declares no listing
+// endpoint. Gateways vary too much to have one, and the caller should fall
+// back to the catalog list plus free-text entry rather than treating this as
+// a failure.
+var ErrNoDiscovery = errors.New("provider has no model-discovery endpoint")
+
+// ErrInvalidRequest marks a discovery failure caused by the caller's own input
+// rather than by the vendor or by this server. Every one of these is reachable
+// from a well-formed request carrying a bad field value, so the handler owes
+// the caller a 400 — a 500 would both misinform them and bury real server
+// faults in the error rate.
+var ErrInvalidRequest = errors.New("invalid discovery request")
+
+// Model is one discovered model.
+type Model struct {
+	// ID is the identifier to register on the provider record, in the form the
+	// vendor issues it. For Bedrock that is the region-prefixed inference
+	// profile id, which is the only form AWS accepts at invoke time.
+	ID string
+	// Label is the vendor's display name where it supplies one.
+	Label string
+	// PricingKnown reports whether the shipped pricing table can price this
+	// model. False means the operator must set rates, or the request would
+	// meter at zero.
+	PricingKnown bool
+	// The rates below are the defaults for this model, taken from the same
+	// table the proxy bills with, so the form prefills exactly what a request
+	// would cost. All zero when PricingKnown is false — an unpriced model is
+	// offered at zero and flagged, rather than withheld: the vendor says the
+	// credential can reach it, and refusing to show it would hide a model the
+	// operator genuinely has.
+	InputPer1k         float64
+	OutputPer1k        float64
+	CachedInputPer1k   float64
+	CacheReadPer1k     float64
+	CacheCreationPer1k float64
+}
+
+// Request identifies which vendor to ask and with what credential.
+type Request struct {
+	// CatalogID selects the catalog entry, which supplies the endpoint, the
+	// auth header and the response shape. The caller never supplies those.
+	CatalogID string
+	// UpstreamURL is the provider record's configured upstream. It is used
+	// only when the catalog entry declares no discovery host of its own.
+	UpstreamURL string
+	// Region substitutes the catalog host's <region> placeholder.
+	Region string
+	// APIKey is the operator's credential, exactly as stored on the record.
+	APIKey string
+}
+
+// Client fetches model listings. The zero value is usable; Resolver and
+// HTTPClient exist so tests can drive it against a local server.
+type Client struct {
+	HTTPClient *http.Client
+	// Resolver looks up the host for the SSRF check. Nil uses the default.
+	Resolver *net.Resolver
+	// AllowPrivateHosts disables the private-address guard. Only tests set it:
+	// their server is on loopback, which is precisely what the guard blocks.
+	AllowPrivateHosts bool
+}
+
+// Fetch returns the models the credential can reach.
+func (c *Client) Fetch(ctx context.Context, req Request) ([]Model, error) {
+	entry, ok := catalog.Lookup(req.CatalogID)
+	if !ok {
+		return nil, fmt.Errorf("%w: unknown catalog provider %q", ErrInvalidRequest, req.CatalogID)
+	}
+	if entry.Discovery == nil {
+		return nil, ErrNoDiscovery
+	}
+
+	endpoint, err := c.discoveryURL(entry, req)
+	if err != nil {
+		return nil, err
+	}
+
+	ctx, cancel := context.WithTimeout(ctx, fetchTimeout)
+	defer cancel()
+
+	httpReq, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
+	if err != nil {
+		return nil, fmt.Errorf("build discovery request: %w", err)
+	}
+	if err := applyAuth(httpReq, entry, req.APIKey); err != nil {
+		return nil, err
+	}
+	for name, value := range entry.Discovery.Headers {
+		httpReq.Header.Set(name, value)
+	}
+	httpReq.Header.Set("Accept", "application/json")
+
+	resp, err := c.httpClient().Do(httpReq)
+	if err != nil {
+		return nil, fmt.Errorf("reach %s: %w", entry.Name, err)
+	}
+	defer func() { _ = resp.Body.Close() }()
+
+	body, err := io.ReadAll(io.LimitReader(resp.Body, maxListingBytes))
+	if err != nil {
+		return nil, fmt.Errorf("read %s listing: %w", entry.Name, err)
+	}
+	if resp.StatusCode != http.StatusOK {
+		// Surface the vendor's own status. An operator whose key lacks a scope
+		// needs to see 403 rather than a generic failure.
+		return nil, fmt.Errorf("%s returned %d for its model listing", entry.Name, resp.StatusCode)
+	}
+
+	ids, err := parseListing(entry.Discovery.Shape, body)
+	if err != nil {
+		return nil, err
+	}
+	return decorate(entry, ids), nil
+}
+
+// discoveryURL builds the listing URL and refuses one that does not point at a
+// public host.
+//
+// The path, query and (for Bedrock) the host all come from the catalog rather
+// than from the caller, so the only operator-controlled part is the host of an
+// entry whose listing lives on its own upstream. That still has to be checked:
+// management holds credentials for every provider, and an upstream pointed at
+// an internal address would turn this endpoint into a probe of the management
+// server's own network.
+func (c *Client) discoveryURL(entry catalog.Provider, req Request) (string, error) {
+	host := entry.Discovery.Host
+	if host == "" {
+		parsed, err := url.Parse(strings.TrimSpace(req.UpstreamURL))
+		if err != nil || parsed.Host == "" {
+			return "", fmt.Errorf("%w: provider upstream %q is not a usable URL", ErrInvalidRequest, req.UpstreamURL)
+		}
+		host = parsed.Host
+	}
+	if strings.Contains(host, catalog.RegionPlaceholder) {
+		region := strings.TrimSpace(req.Region)
+		if region == "" {
+			// A provider record carries no region field: the region lives
+			// inside the upstream host the operator already configured, so
+			// read it back out rather than asking them for it twice.
+			region = RegionFromUpstream(entry, req.UpstreamURL)
+		}
+		if region == "" {
+			return "", fmt.Errorf("%w: %s discovery needs a region, and none could be read from the provider upstream",
+				ErrInvalidRequest, entry.Name)
+		}
+		host = strings.ReplaceAll(host, catalog.RegionPlaceholder, region)
+	}
+
+	target := &url.URL{Scheme: "https", Host: host, Path: entry.Discovery.Path, RawQuery: entry.Discovery.Query}
+	if err := c.checkPublicHost(target.Hostname()); err != nil {
+		return "", err
+	}
+	return target.String(), nil
+}
+
+// RegionFromUpstream recovers the region an operator embedded in the provider
+// upstream, by matching it against the catalog's own host template. Bedrock's
+// template is "bedrock-runtime.<region>.amazonaws.com" and Vertex's is
+// "<region>-aiplatform.googleapis.com", so the region is whatever sits between
+// the fixed halves. Returns empty when the upstream does not match the
+// template, which is the case for a custom or proxied endpoint.
+func RegionFromUpstream(entry catalog.Provider, upstreamURL string) string {
+	prefix, suffix, found := strings.Cut(entry.DefaultHost, catalog.RegionPlaceholder)
+	if !found {
+		return ""
+	}
+	parsed, err := url.Parse(strings.TrimSpace(upstreamURL))
+	if err != nil {
+		return ""
+	}
+	host := parsed.Hostname()
+	if host == "" {
+		// A bare host with no scheme parses as a path, not a host.
+		host = strings.TrimSpace(upstreamURL)
+	}
+	// The two halves must not overlap. "bedrock-runtime.amazonaws.com" carries
+	// both of Bedrock's — it is the regionless endpoint — and satisfies both
+	// checks above while leaving nothing between them, so slicing it would
+	// panic on an inverted range rather than report "no region here".
+	if !strings.HasPrefix(host, prefix) || !strings.HasSuffix(host, suffix) ||
+		len(host) < len(prefix)+len(suffix) {
+		return ""
+	}
+	region := host[len(prefix) : len(host)-len(suffix)]
+	if region == "" || strings.Contains(region, ".") {
+		return ""
+	}
+	return region
+}
+
+// checkPublicHost refuses hosts that resolve to an address the management
+// server should never be asked to reach on an operator's behalf.
+func (c *Client) checkPublicHost(host string) error {
+	if c.AllowPrivateHosts {
+		return nil
+	}
+	if host == "" {
+		return errors.New("discovery host is empty")
+	}
+	resolver := c.Resolver
+	if resolver == nil {
+		resolver = net.DefaultResolver
+	}
+	ctx, cancel := context.WithTimeout(context.Background(), fetchTimeout)
+	defer cancel()
+
+	addrs, err := resolver.LookupNetIP(ctx, "ip", host)
+	if err != nil {
+		return fmt.Errorf("resolve discovery host %q: %w", host, err)
+	}
+	// Every address must be public: a name that resolves to one public and one
+	// loopback address is still a way to reach loopback.
+	for _, addr := range addrs {
+		if !isPublic(addr) {
+			return fmt.Errorf("%w: discovery host %q resolves to a non-public address", ErrInvalidRequest, host)
+		}
+	}
+	return nil
+}
+
+// isPublic reports whether an address is one we are willing to dial.
+func isPublic(addr netip.Addr) bool {
+	addr = addr.Unmap()
+	switch {
+	case !addr.IsValid(),
+		addr.IsLoopback(),
+		addr.IsPrivate(),
+		addr.IsLinkLocalUnicast(),
+		addr.IsLinkLocalMulticast(),
+		addr.IsInterfaceLocalMulticast(),
+		addr.IsMulticast(),
+		addr.IsUnspecified():
+		return false
+	}
+	// 100.64.0.0/10 (carrier NAT) is where NetBird's own overlay addresses
+	// live, so it is emphatically not somewhere to send a provider credential.
+	if addr.Is4() {
+		b := addr.As4()
+		if b[0] == 100 && b[1] >= 64 && b[1] <= 127 {
+			return false
+		}
+	}
+	return true
+}
+
+// applyAuth sets the credential header the catalog entry declares. A Vertex
+// service-account key is exchanged for an OAuth token first, the same way the
+// proxy does at request time.
+func applyAuth(req *http.Request, entry catalog.Provider, apiKey string) error {
+	key := strings.TrimSpace(apiKey)
+	if key == "" {
+		return fmt.Errorf("%w: %s discovery needs an API key", ErrInvalidRequest, entry.Name)
+	}
+	if rest, ok := strings.CutPrefix(key, vertexKeyfilePrefix); ok {
+		token, err := mintGCPToken(req.Context(), rest)
+		if err != nil {
+			return err
+		}
+		key = token
+	}
+	name := entry.AuthHeaderName
+	if name == "" {
+		name = "Authorization"
+	}
+	template := entry.AuthHeaderTemplate
+	if template == "" {
+		template = "${API_KEY}"
+	}
+	req.Header.Set(name, strings.ReplaceAll(template, "${API_KEY}", key))
+	return nil
+}
+
+// mintGCPToken exchanges a base64 service-account key for an access token.
+func mintGCPToken(ctx context.Context, saKeyB64 string) (string, error) {
+	jsonKey, err := base64.StdEncoding.DecodeString(strings.TrimSpace(saKeyB64))
+	if err != nil {
+		return "", fmt.Errorf("decode service-account key: %w", err)
+	}
+	conf, err := google.JWTConfigFromJSON(jsonKey, gcpScope)
+	if err != nil {
+		return "", fmt.Errorf("parse service-account key: %w", err)
+	}
+	tok, err := conf.TokenSource(ctx).Token()
+	if err != nil {
+		return "", fmt.Errorf("mint gcp token: %w", err)
+	}
+	return tok.AccessToken, nil
+}
+
+// decorate turns raw vendor ids into the models the caller renders, attaching
+// the rates the request would actually be billed at.
+//
+// Rates come from the live default pricing table rather than the compiled-in
+// catalog, because that is the table the synthesiser ships to the proxy: an
+// operator running a defaults_llm_pricing.yaml would otherwise be shown one
+// price in the form and charged another. It is also the same lookup the catalog
+// endpoint prefills from, so a model reached by either route prices identically.
+func decorate(entry catalog.Provider, ids []listedModel) []Model {
+	out := make([]Model, 0, len(ids))
+	seen := make(map[string]struct{}, len(ids))
+	for _, listed := range ids {
+		if listed.id == "" {
+			continue
+		}
+		if _, dup := seen[listed.id]; dup {
+			continue
+		}
+		seen[listed.id] = struct{}{}
+
+		// The table keys pricing by the normalised id while the vendor issues
+		// the wire form, so normalise before looking it up — otherwise every
+		// Bedrock profile would report unpriced.
+		model := Model{ID: listed.id, Label: listed.label}
+		if rate, known := pricing.LookupDefault(entry.PricingSurfaces, normalizeForPricing(entry.ID, listed.id)); known {
+			model.PricingKnown = true
+			model.InputPer1k = rate.InputPer1k
+			model.OutputPer1k = rate.OutputPer1k
+			model.CachedInputPer1k = rate.CachedInputPer1k
+			model.CacheReadPer1k = rate.CacheReadPer1k
+			model.CacheCreationPer1k = rate.CacheCreationPer1k
+		}
+		out = append(out, model)
+	}
+	return out
+}
+
+// refuseRedirect is the redirect policy every discovery request runs under. A
+// redirect is a way to move the request to a host checkPublicHost never saw,
+// so none are followed.
+func refuseRedirect(*http.Request, []*http.Request) error {
+	return http.ErrUseLastResponse
+}
+
+func (c *Client) httpClient() *http.Client {
+	if c.HTTPClient != nil {
+		if c.HTTPClient.CheckRedirect != nil {
+			return c.HTTPClient
+		}
+		// An injected client that states no policy still gets ours: the
+		// no-redirect guarantee should not depend on the caller remembering it.
+		//
+		// Copied rather than assigned into: one Client is shared by every
+		// request for the process's lifetime, so writing to its fields here
+		// would race across request goroutines. The copy shares the Transport,
+		// which is safe for concurrent use by design.
+		clone := *c.HTTPClient
+		clone.CheckRedirect = refuseRedirect
+		return &clone
+	}
+	transport := guardedTransport
+	if c.AllowPrivateHosts {
+		transport = http.DefaultTransport
+	}
+	return &http.Client{
+		Timeout:       fetchTimeout,
+		Transport:     transport,
+		CheckRedirect: refuseRedirect,
+	}
+}
+
+// guardedTransport dials only addresses isPublic accepts.
+//
+// checkPublicHost resolves the host itself, and the transport then resolves it
+// again when it dials — two lookups of a name whose owner chooses the answers.
+// A record that returns a public address to the first and 127.0.0.1 to the
+// second passes the guard and reaches loopback anyway, which is the whole of
+// DNS rebinding. Re-checking at the socket closes that window: whatever the
+// second lookup returned is what Control is handed, and an address the guard
+// refuses never gets connected.
+//
+// Shared package-wide rather than built per Fetch so connections and their
+// pool survive between calls; the guard holds no state.
+var guardedTransport = newGuardedTransport()
+
+func newGuardedTransport() http.RoundTripper {
+	base, ok := http.DefaultTransport.(*http.Transport)
+	if !ok {
+		// Something replaced the default transport. Fall back to it rather
+		// than dropping its behaviour, and rely on checkPublicHost alone.
+		return http.DefaultTransport
+	}
+	// Cloned so proxy settings, TLS defaults and timeouts come from the
+	// standard transport rather than being restated here.
+	transport := base.Clone()
+	dialer := &net.Dialer{
+		Timeout:   fetchTimeout,
+		KeepAlive: 30 * time.Second,
+		Control: func(_, address string, _ syscall.RawConn) error {
+			return guardDialAddress(address)
+		},
+	}
+	transport.DialContext = dialer.DialContext
+	return transport
+}
+
+// guardDialAddress refuses a resolved socket address the discovery client has
+// no business connecting to. Control hands it over post-resolution and
+// pre-connect, once per address the dialer tries, so a name with several A
+// records is checked at each one.
+func guardDialAddress(address string) error {
+	host, _, err := net.SplitHostPort(address)
+	if err != nil {
+		return fmt.Errorf("discovery dial address %q is unreadable", address)
+	}
+	addr, err := netip.ParseAddr(host)
+	if err != nil {
+		// Control is documented to receive a resolved address; anything else
+		// is a state we cannot vet, so it does not get dialled.
+		return fmt.Errorf("discovery dial address %q is not an IP", host)
+	}
+	if !isPublic(addr) {
+		return fmt.Errorf("discovery refused to dial non-public address %s", addr)
+	}
+	return nil
+}

management/internals/modules/agentnetwork/modeldiscovery/discovery_test.go

@@ -0,0 +1,532 @@
+package modeldiscovery
+
+import (
+	"context"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"net/netip"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/catalog"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/pricing"
+)
+
+// stubTransport answers every request with one canned response and records the
+// request it was given, so a test can assert on the URL and headers the client
+// built without a network round trip.
+type stubTransport struct {
+	status int
+	body   string
+	got    *http.Request
+}
+
+func (s *stubTransport) RoundTrip(req *http.Request) (*http.Response, error) {
+	s.got = req
+	status := s.status
+	if status == 0 {
+		status = http.StatusOK
+	}
+	return &http.Response{
+		StatusCode: status,
+		Body:       io.NopCloser(strings.NewReader(s.body)),
+		Header:     http.Header{"Content-Type": []string{"application/json"}},
+		Request:    req,
+	}, nil
+}
+
+// newStubClient returns a client that never leaves the process. The host guard
+// is disabled because it would otherwise resolve the vendor's real name, which
+// would make these tests depend on DNS.
+func newStubClient(status int, body string) (*Client, *stubTransport) {
+	tr := &stubTransport{status: status, body: body}
+	return &Client{
+		HTTPClient:        &http.Client{Transport: tr},
+		AllowPrivateHosts: true,
+	}, tr
+}
+
+// The payloads below are trimmed from what the vendors actually returned in
+// the discovery e2e, rather than invented, so a parser that only works against
+// an idealised shape fails here.
+
+const openAIListing = `{"object":"list","data":[
+  {"id":"gpt-4o-mini","object":"model","created":1721172741,"owned_by":"system"},
+  {"id":"gpt-4o","object":"model","created":1715367049,"owned_by":"system"}
+]}`
+
+const anthropicListing = `{"data":[
+  {"type":"model","id":"claude-haiku-4-5-20251001","display_name":"Claude Haiku 4.5"},
+  {"type":"model","id":"claude-sonnet-4-6","display_name":"Claude Sonnet 4.6"}
+],"has_more":false}`
+
+const bedrockListing = `{"inferenceProfileSummaries":[
+  {"inferenceProfileId":"eu.anthropic.claude-haiku-4-5-20251001-v1:0",
+   "inferenceProfileName":"EU Anthropic Claude Haiku 4.5","status":"ACTIVE","type":"SYSTEM_DEFINED"},
+  {"inferenceProfileId":"global.cohere.embed-v4:0",
+   "inferenceProfileName":"Global Cohere Embed v4","status":"ACTIVE","type":"SYSTEM_DEFINED"},
+  {"inferenceProfileId":"eu.meta.llama3-2-1b-instruct-v1:0",
+   "inferenceProfileName":"EU Meta Llama 3.2 1B","status":"INACTIVE","type":"SYSTEM_DEFINED"}
+]}`
+
+const vertexListing = `{"publisherModels":[
+  {"name":"publishers/anthropic/models/claude-3-opus","versionId":"20240229","launchStage":"GA"},
+  {"name":"publishers/anthropic/models/claude-sonnet-4-5","versionId":"20250929","launchStage":"GA"}
+]}`
+
+func TestFetchOpenAIListing(t *testing.T) {
+	cl, tr := newStubClient(http.StatusOK, openAIListing)
+
+	models, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "openai_api",
+		UpstreamURL: "https://api.openai.com",
+		APIKey:      "sk-test",
+	})
+	require.NoError(t, err)
+
+	assert.Equal(t, "https://api.openai.com/v1/models", tr.got.URL.String())
+	assert.Equal(t, "Bearer sk-test", tr.got.Header.Get("Authorization"),
+		"the credential must be injected through the catalog's auth template")
+	assert.Equal(t, []string{"gpt-4o-mini", "gpt-4o"}, ids(models))
+	for _, m := range models {
+		assert.True(t, m.PricingKnown, "both models are in the shipped catalog: %s", m.ID)
+	}
+}
+
+func TestFetchAnthropicSendsTheVersionHeader(t *testing.T) {
+	cl, tr := newStubClient(http.StatusOK, anthropicListing)
+
+	models, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "anthropic_api",
+		UpstreamURL: "https://api.anthropic.com",
+		APIKey:      "sk-ant-test",
+	})
+	require.NoError(t, err)
+
+	// Anthropic rejects a request without the version header, so a listing
+	// that reached us at all proves it was sent — but assert it, because the
+	// failure mode otherwise only shows up against the live API.
+	assert.Equal(t, "2023-06-01", tr.got.Header.Get("anthropic-version"))
+	assert.Equal(t, "sk-ant-test", tr.got.Header.Get("x-api-key"),
+		"Anthropic takes a bare key under its own header, not a Bearer token")
+	assert.Equal(t, "limit=1000", tr.got.URL.RawQuery)
+
+	assert.Equal(t, []string{"claude-haiku-4-5-20251001", "claude-sonnet-4-6"}, ids(models))
+	assert.Equal(t, "Claude Haiku 4.5", models[0].Label)
+}
+
+func TestFetchBedrockUsesTheControlPlaneAndKeepsWireIDs(t *testing.T) {
+	cl, tr := newStubClient(http.StatusOK, bedrockListing)
+
+	models, err := cl.Fetch(context.Background(), Request{
+		CatalogID: "bedrock_api",
+		// The record's upstream is the RUNTIME host, which does not serve
+		// listings. The catalog's own discovery host must win over it.
+		UpstreamURL: "https://bedrock-runtime.eu-central-1.amazonaws.com",
+		Region:      "eu-central-1",
+		APIKey:      "aws-bearer",
+	})
+	require.NoError(t, err)
+
+	assert.Equal(t, "https://bedrock.eu-central-1.amazonaws.com/inference-profiles",
+		tr.got.URL.String(), "listings come from the control plane, not the runtime host")
+
+	// Region-prefixed ids verbatim: the prefix is what makes them invocable
+	// and it cannot be reconstructed — global.* alongside eu.* is exactly the
+	// case that defeats deriving it from the configured region.
+	assert.Equal(t, []string{
+		"eu.anthropic.claude-haiku-4-5-20251001-v1:0",
+		"global.cohere.embed-v4:0",
+	}, ids(models), "an INACTIVE profile must not be offered")
+
+	assert.True(t, models[0].PricingKnown,
+		"the catalog prices anthropic.claude-haiku-4-5, which this id normalises to")
+	assert.False(t, models[1].PricingKnown,
+		"cohere embed is not in the shipped Bedrock catalog, so the operator must price it")
+
+	// The rates travel with the model, so the form can prefill an editable row
+	// rather than making the operator look every price up by hand.
+	assert.Positive(t, models[0].InputPer1k, "a priced model must carry its input rate")
+	assert.Positive(t, models[0].OutputPer1k, "a priced model must carry its output rate")
+	// An unpriced model is offered at zero and flagged, not withheld: the
+	// vendor says the credential can reach it.
+	assert.Zero(t, models[1].InputPer1k)
+	assert.Zero(t, models[1].OutputPer1k)
+}
+
+// TestDiscoveredRatesMatchTheCatalogEndpoint pins the two prefill paths to one
+// table. The provider form fills a model row either from the catalog response
+// or from a discovery response, and an operator who switches between them must
+// not see the price change — both must equal what the proxy will bill.
+func TestDiscoveredRatesMatchTheCatalogEndpoint(t *testing.T) {
+	cl, _ := newStubClient(http.StatusOK, openAIListing)
+
+	models, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "openai_api",
+		UpstreamURL: "https://api.openai.com",
+		APIKey:      "sk-test",
+	})
+	require.NoError(t, err)
+	require.NotEmpty(t, models)
+
+	entry, ok := catalog.Lookup("openai_api")
+	require.True(t, ok)
+
+	for _, m := range models {
+		want, known := pricing.LookupDefault(entry.PricingSurfaces, m.ID)
+		require.True(t, known, "%s should be priced by the default table", m.ID)
+		assert.Equal(t, want.InputPer1k, m.InputPer1k, "input rate for %s", m.ID)
+		assert.Equal(t, want.OutputPer1k, m.OutputPer1k, "output rate for %s", m.ID)
+		assert.Equal(t, want.CachedInputPer1k, m.CachedInputPer1k, "cached-input rate for %s", m.ID)
+	}
+}
+
+func TestFetchVertexJoinsNameAndVersion(t *testing.T) {
+	cl, _ := newStubClient(http.StatusOK, vertexListing)
+
+	models, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "vertex_ai_api",
+		UpstreamURL: "https://us-east5-aiplatform.googleapis.com",
+		Region:      "us-east5",
+		APIKey:      "ya29.test-token",
+	})
+	require.NoError(t, err)
+
+	// Vertex addresses a model as "<id>@<version>" on rawPredict, and splits
+	// those across two fields in the listing.
+	assert.Equal(t, []string{"claude-3-opus@20240229", "claude-sonnet-4-5@20250929"}, ids(models))
+	assert.Equal(t, "claude-3-opus", models[0].Label)
+}
+
+func TestFetchSurfacesTheVendorStatus(t *testing.T) {
+	cl, _ := newStubClient(http.StatusForbidden, `{"error":{"message":"no access"}}`)
+
+	_, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "openai_api",
+		UpstreamURL: "https://api.openai.com",
+		APIKey:      "sk-test",
+	})
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "403",
+		"an operator whose key lacks access needs to see which status the vendor returned")
+}
+
+func TestFetchRejectsAProviderWithoutDiscovery(t *testing.T) {
+	cl, _ := newStubClient(http.StatusOK, openAIListing)
+
+	_, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "litellm_proxy",
+		UpstreamURL: "https://gateway.example.com",
+		APIKey:      "sk-test",
+	})
+	assert.ErrorIs(t, err, ErrNoDiscovery,
+		"a gateway with no listing endpoint must be distinguishable from a failure, so the caller can fall back")
+}
+
+func TestFetchRequiresACredential(t *testing.T) {
+	cl, _ := newStubClient(http.StatusOK, openAIListing)
+
+	_, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "openai_api",
+		UpstreamURL: "https://api.openai.com",
+	})
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "API key")
+}
+
+func TestDiscoveryURLNeedsARegionWhenTheHostTemplatesOne(t *testing.T) {
+	cl, _ := newStubClient(http.StatusOK, bedrockListing)
+
+	// An upstream that matches no catalog template — a proxy in front of
+	// Bedrock, say — leaves nothing to read the region from. Refusing beats
+	// guessing: an unsubstituted placeholder would dial a host that does not
+	// exist, and a guessed region would dial the wrong account's endpoint.
+	_, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "bedrock_api",
+		UpstreamURL: "https://bedrock.internal-proxy.example.com",
+		APIKey:      "aws-bearer",
+	})
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "region")
+}
+
+// TestHostGuardRejectsNonPublicAddresses is the SSRF guard. Management holds a
+// credential for every provider, so an upstream pointed at an internal address
+// would turn discovery into a way to probe — and hand a token to — the
+// management server's own network.
+func TestHostGuardRejectsNonPublicAddresses(t *testing.T) {
+	for _, tc := range []struct {
+		name string
+		addr string
+		want bool
+	}{
+		{"loopback v4", "127.0.0.1", false},
+		{"loopback v6", "::1", false},
+		{"private 10/8", "10.0.0.5", false},
+		{"private 172.16/12", "172.16.4.1", false},
+		{"private 192.168/16", "192.168.1.1", false},
+		{"link-local", "169.254.169.254", false}, // cloud metadata
+		{"unspecified", "0.0.0.0", false},
+		{"multicast", "224.0.0.1", false},
+		{"netbird overlay 100.64/10", "100.90.1.2", false},
+		{"v4-mapped loopback", "::ffff:127.0.0.1", false},
+		{"public v4", "1.1.1.1", true},
+		{"public v6", "2606:4700:4700::1111", true},
+		{"just outside CGNAT", "100.128.0.1", true},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			addr, err := netip.ParseAddr(tc.addr)
+			require.NoError(t, err)
+			assert.Equal(t, tc.want, isPublic(addr))
+		})
+	}
+}
+
+func TestHostGuardResolvesAndRejectsLocalhost(t *testing.T) {
+	cl := &Client{}
+	err := cl.checkPublicHost("localhost")
+	require.Error(t, err, "a name resolving to loopback must be refused, not just a literal address")
+	assert.Contains(t, err.Error(), "non-public")
+}
+
+// TestRedirectsAreNotFollowed covers a gap the other tests leave open: they all
+// inject an HTTPClient, which bypasses httpClient() and therefore the redirect
+// policy entirely. The policy is a security control — a 302 moves the request
+// to a host checkPublicHost never resolved — so it needs a test that goes
+// through the constructor the manager actually uses.
+func TestRedirectsAreNotFollowed(t *testing.T) {
+	var hits int
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		hits++
+		http.Redirect(w, r, "http://169.254.169.254/latest/meta-data/", http.StatusFound)
+	}))
+	t.Cleanup(srv.Close)
+
+	for name, cl := range map[string]*Client{
+		// The production shape: no injected client at all.
+		"default client": {AllowPrivateHosts: true},
+		// An injected client that states no policy must inherit ours rather
+		// than silently chasing the redirect.
+		"injected client with no policy": {
+			AllowPrivateHosts: true,
+			HTTPClient:        &http.Client{},
+		},
+	} {
+		t.Run(name, func(t *testing.T) {
+			hits = 0
+			req, err := http.NewRequest(http.MethodGet, srv.URL, nil)
+			require.NoError(t, err)
+
+			resp, err := cl.httpClient().Do(req)
+			require.NoError(t, err)
+			t.Cleanup(func() { _ = resp.Body.Close() })
+
+			assert.Equal(t, http.StatusFound, resp.StatusCode,
+				"the redirect must be surfaced, not followed to an unchecked host")
+			assert.Equal(t, 1, hits, "exactly one request must leave the client")
+		})
+	}
+}
+
+// TestInjectedClientKeepsItsOwnRedirectPolicy pins that the default above is a
+// default, not an override, and that supplying it does not mutate the caller's
+// client — one Client is shared across every request, so a write here would
+// race.
+func TestInjectedClientKeepsItsOwnRedirectPolicy(t *testing.T) {
+	own := func(*http.Request, []*http.Request) error { return nil }
+	injected := &http.Client{CheckRedirect: own}
+	cl := &Client{HTTPClient: injected}
+
+	assert.Same(t, injected, cl.httpClient(),
+		"a client that states a policy must be handed back untouched")
+
+	bare := &http.Client{}
+	cl = &Client{HTTPClient: bare}
+	require.NotSame(t, bare, cl.httpClient(), "the policy must be applied to a copy")
+	assert.Nil(t, bare.CheckRedirect, "the caller's client must not be written to")
+}
+
+// TestDialGuardRejectsRebindingToANonPublicAddress covers the window between
+// the two DNS lookups. checkPublicHost resolves the host, then the transport
+// resolves it again to dial; a name whose owner answers the first with a public
+// address and the second with 127.0.0.1 would otherwise pass the guard and
+// still reach loopback. The dial-time check sees whatever the second lookup
+// actually returned.
+func TestDialGuardRejectsRebindingToANonPublicAddress(t *testing.T) {
+	for _, tc := range []struct {
+		name    string
+		address string
+		wantErr string
+	}{
+		{"loopback", "127.0.0.1:443", "non-public"},
+		{"cloud metadata", "169.254.169.254:80", "non-public"},
+		{"rfc1918", "10.1.2.3:443", "non-public"},
+		{"netbird overlay", "100.90.1.2:443", "non-public"},
+		{"loopback v6", "[::1]:443", "non-public"},
+		{"unresolved name", "evil.example.com:443", "not an IP"},
+		{"no port", "1.1.1.1", "unreadable"},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			err := guardDialAddress(tc.address)
+			require.Error(t, err)
+			assert.Contains(t, err.Error(), tc.wantErr)
+		})
+	}
+
+	assert.NoError(t, guardDialAddress("1.1.1.1:443"), "a public address must still be dialled")
+	assert.NoError(t, guardDialAddress("[2606:4700:4700::1111]:443"))
+}
+
+// TestDialGuardIsInstalledOnTheDefaultClient pins the wiring rather than the
+// guard: a correct guard nothing calls protects nothing.
+func TestDialGuardIsInstalledOnTheDefaultClient(t *testing.T) {
+	cl := &Client{}
+	transport, ok := cl.httpClient().Transport.(*http.Transport)
+	require.True(t, ok, "the default discovery client must carry the guarded transport")
+	require.NotNil(t, transport.DialContext, "the guarded transport must dial through the guard")
+
+	_, err := transport.DialContext(context.Background(), "tcp", "127.0.0.1:9")
+	require.Error(t, err, "the guard must refuse loopback even when the caller dials it directly")
+	assert.Contains(t, err.Error(), "non-public")
+
+	// Tests point the client at a loopback server on purpose, so the opt-out
+	// has to reach the dialer too.
+	relaxed := &Client{AllowPrivateHosts: true}
+	assert.Equal(t, http.DefaultTransport, relaxed.httpClient().Transport)
+}
+
+// TestCallerInputFailuresAreMarkedInvalid keeps the handler's 400 mapping
+// honest: it branches on this sentinel, so an unmarked caller-input failure
+// silently becomes a 500.
+func TestCallerInputFailuresAreMarkedInvalid(t *testing.T) {
+	for _, tc := range []struct {
+		name string
+		req  Request
+	}{
+		{"unknown provider", Request{CatalogID: "not_a_provider", APIKey: "k"}},
+		{"unusable upstream", Request{CatalogID: "openai_api", UpstreamURL: "://", APIKey: "k"}},
+		{"missing api key", Request{CatalogID: "openai_api", UpstreamURL: "https://api.openai.com"}},
+		{"no region to read", Request{
+			CatalogID:   "bedrock_api",
+			UpstreamURL: "https://bedrock-runtime.amazonaws.com",
+			APIKey:      "aws-bearer",
+		}},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			cl, _ := newStubClient(http.StatusOK, openAIListing)
+			_, err := cl.Fetch(context.Background(), tc.req)
+			require.Error(t, err)
+			assert.ErrorIs(t, err, ErrInvalidRequest)
+		})
+	}
+}
+
+// TestEveryDiscoveryEntryHasAParser keeps the catalog and the parser table from
+// drifting: adding a Discovery block with a shape nothing parses would fail
+// only at runtime, in front of an operator.
+func TestEveryDiscoveryEntryHasAParser(t *testing.T) {
+	for _, entry := range catalog.All() {
+		if entry.Discovery == nil {
+			continue
+		}
+		t.Run(entry.ID, func(t *testing.T) {
+			assert.NotEmpty(t, entry.Discovery.Path, "a discovery entry needs a path")
+			_, err := parseListing(entry.Discovery.Shape, []byte(`{}`))
+			assert.NoError(t, err, "shape %q has no parser", entry.Discovery.Shape)
+		})
+	}
+}
+
+func ids(models []Model) []string {
+	out := make([]string, 0, len(models))
+	for _, m := range models {
+		out = append(out, m.ID)
+	}
+	return out
+}
+
+// TestRegionIsReadBackFromTheUpstream covers the reason the API takes no
+// region field: a provider record has none, and the operator already encoded
+// it in the upstream host when they configured inference.
+func TestRegionIsReadBackFromTheUpstream(t *testing.T) {
+	cl, tr := newStubClient(http.StatusOK, bedrockListing)
+
+	_, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "bedrock_api",
+		UpstreamURL: "https://bedrock-runtime.us-west-2.amazonaws.com",
+		APIKey:      "aws-bearer",
+	})
+	require.NoError(t, err)
+	assert.Equal(t, "bedrock.us-west-2.amazonaws.com", tr.got.URL.Host)
+}
+
+func TestRegionFromUpstream(t *testing.T) {
+	bedrock, ok := catalog.Lookup("bedrock_api")
+	require.True(t, ok)
+	vertex, ok := catalog.Lookup("vertex_ai_api")
+	require.True(t, ok)
+
+	for _, tc := range []struct {
+		name     string
+		entry    catalog.Provider
+		upstream string
+		want     string
+	}{
+		{"bedrock runtime host", bedrock, "https://bedrock-runtime.eu-central-1.amazonaws.com", "eu-central-1"},
+		{"bedrock without scheme", bedrock, "bedrock-runtime.ap-south-1.amazonaws.com", "ap-south-1"},
+		{"vertex regional host", vertex, "https://us-east5-aiplatform.googleapis.com", "us-east5"},
+		// A proxied or self-hosted upstream matches no template, and guessing
+		// a region from it would build a URL pointing somewhere arbitrary.
+		{"unrelated upstream", bedrock, "https://llm.internal.example.com", ""},
+		{"vertex global host has no region segment", vertex, "https://aiplatform.googleapis.com", ""},
+		// Bedrock's regionless endpoint carries both halves of the template at
+		// once, with nothing between them. It has to read as "no region here"
+		// rather than as an inverted slice range.
+		{"bedrock regionless endpoint", bedrock, "https://bedrock-runtime.amazonaws.com", ""},
+		{"bedrock regionless without scheme", bedrock, "bedrock-runtime.amazonaws.com", ""},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			assert.Equal(t, tc.want, RegionFromUpstream(tc.entry, tc.upstream))
+		})
+	}
+}
+
+// bedrockGeoListing carries profiles from geographies the original prefix list
+// did not name. Every one reduces to a catalog key, so every one must arrive
+// priced — an unstripped geography is what made a real account's listing come
+// back almost entirely at zero.
+const bedrockGeoListing = `{"inferenceProfileSummaries":[
+  {"inferenceProfileId":"jp.anthropic.claude-sonnet-5-20260514-v1:0",
+   "inferenceProfileName":"JP Anthropic Claude Sonnet 5","status":"ACTIVE","type":"SYSTEM_DEFINED"},
+  {"inferenceProfileId":"au.anthropic.claude-haiku-4-5-20251001-v1:0",
+   "inferenceProfileName":"AU Anthropic Claude Haiku 4.5","status":"ACTIVE","type":"SYSTEM_DEFINED"},
+  {"inferenceProfileId":"us-gov.anthropic.claude-sonnet-5-20260514-v1:0",
+   "inferenceProfileName":"GovCloud Anthropic Claude Sonnet 5","status":"ACTIVE","type":"SYSTEM_DEFINED"}
+]}`
+
+func TestBedrockProfilesFromAnyGeographyArrivePriced(t *testing.T) {
+	cl, _ := newStubClient(http.StatusOK, bedrockGeoListing)
+
+	models, err := cl.Fetch(context.Background(), Request{
+		CatalogID:   "bedrock_api",
+		UpstreamURL: "https://bedrock-runtime.eu-central-1.amazonaws.com",
+		APIKey:      "aws-token",
+	})
+	require.NoError(t, err)
+	require.Len(t, models, 3)
+
+	for _, m := range models {
+		assert.True(t, m.PricingKnown, "%s must resolve to a catalog rate", m.ID)
+		assert.Greater(t, m.InputPer1k, 0.0, "input rate for %s", m.ID)
+		assert.Greater(t, m.OutputPer1k, 0.0, "output rate for %s", m.ID)
+		assert.Greater(t, m.CacheReadPer1k, 0.0, "cache-read rate for %s", m.ID)
+	}
+
+	// The wire id is preserved whatever the pricing key reduced to: it is the
+	// only form that works at invoke time.
+	assert.Equal(t, "jp.anthropic.claude-sonnet-5-20260514-v1:0", models[0].ID)
+}

management/internals/modules/agentnetwork/modeldiscovery/parse.go

@@ -0,0 +1,134 @@
+package modeldiscovery
+
+import (
+	"encoding/json"
+	"fmt"
+	"strings"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/catalog"
+	sharedllm "github.com/netbirdio/netbird/shared/llm"
+)
+
+// listedModel is one entry lifted out of a vendor listing before the catalog
+// is consulted about it.
+type listedModel struct {
+	id    string
+	label string
+}
+
+// parseListing extracts model ids from a vendor listing. Each vendor invented
+// its own envelope, and the shape is declared by the catalog rather than
+// sniffed, so a vendor that changes shape fails loudly instead of silently
+// returning nothing.
+func parseListing(shape catalog.ListingShape, body []byte) ([]listedModel, error) {
+	switch shape {
+	case catalog.ShapeOpenAIData:
+		return parseOpenAIData(body)
+	case catalog.ShapeBedrockInferenceProfiles:
+		return parseBedrockInferenceProfiles(body)
+	case catalog.ShapeVertexPublisherModels:
+		return parseVertexPublisherModels(body)
+	default:
+		return nil, fmt.Errorf("no parser for listing shape %q", shape)
+	}
+}
+
+// parseOpenAIData reads {"data":[{"id":…}]}, which OpenAI defined and
+// Anthropic adopted. Anthropic additionally supplies display_name.
+func parseOpenAIData(body []byte) ([]listedModel, error) {
+	var doc struct {
+		Data []struct {
+			ID          string `json:"id"`
+			DisplayName string `json:"display_name"`
+		} `json:"data"`
+	}
+	if err := json.Unmarshal(body, &doc); err != nil {
+		return nil, fmt.Errorf("decode model listing: %w", err)
+	}
+	out := make([]listedModel, 0, len(doc.Data))
+	for _, entry := range doc.Data {
+		out = append(out, listedModel{id: entry.ID, label: entry.DisplayName})
+	}
+	return out, nil
+}
+
+// parseBedrockInferenceProfiles reads
+// {"inferenceProfileSummaries":[{"inferenceProfileId":…}]}.
+//
+// The profile id is taken verbatim because its region prefix (eu., us.,
+// global.) is what makes it invocable, and it is not derivable from the
+// configured region — an account in one region legitimately holds global.*
+// profiles alongside its regional ones.
+//
+// Only ACTIVE profiles are offered: AWS reports others, and registering one
+// would produce a model that routes inside NetBird and fails at AWS.
+func parseBedrockInferenceProfiles(body []byte) ([]listedModel, error) {
+	var doc struct {
+		Summaries []struct {
+			ID     string `json:"inferenceProfileId"`
+			Name   string `json:"inferenceProfileName"`
+			Status string `json:"status"`
+		} `json:"inferenceProfileSummaries"`
+	}
+	if err := json.Unmarshal(body, &doc); err != nil {
+		return nil, fmt.Errorf("decode inference-profile listing: %w", err)
+	}
+	out := make([]listedModel, 0, len(doc.Summaries))
+	for _, entry := range doc.Summaries {
+		if entry.Status != "" && !strings.EqualFold(entry.Status, "ACTIVE") {
+			continue
+		}
+		out = append(out, listedModel{id: entry.ID, label: entry.Name})
+	}
+	return out, nil
+}
+
+// parseVertexPublisherModels reads {"publisherModels":[{"name":…}]}, where
+// name is a resource path ("publishers/anthropic/models/claude-3-opus") and
+// the version lives in a separate field.
+//
+// Vertex addresses a model as "<id>@<version>" on the rawPredict path, so the
+// two are joined here: reporting the bare name would hand the operator an id
+// that looks usable and is not.
+func parseVertexPublisherModels(body []byte) ([]listedModel, error) {
+	var doc struct {
+		Models []struct {
+			Name      string `json:"name"`
+			VersionID string `json:"versionId"`
+		} `json:"publisherModels"`
+	}
+	if err := json.Unmarshal(body, &doc); err != nil {
+		return nil, fmt.Errorf("decode publisher-model listing: %w", err)
+	}
+	out := make([]listedModel, 0, len(doc.Models))
+	for _, entry := range doc.Models {
+		id := entry.Name
+		if slash := strings.LastIndex(id, "/"); slash >= 0 {
+			id = id[slash+1:]
+		}
+		if id == "" {
+			continue
+		}
+		label := id
+		if entry.VersionID != "" {
+			id += "@" + entry.VersionID
+		}
+		out = append(out, listedModel{id: id, label: label})
+	}
+	return out, nil
+}
+
+// normalizeForPricing maps a vendor's wire id onto the key the catalog prices
+// it under. It mirrors the synthesiser's normalizePricingModelID: the two must
+// agree, or a model reported here as priced would meter at the default rate
+// instead of the operator's.
+func normalizeForPricing(catalogProviderID, modelID string) string {
+	switch {
+	case catalog.IsBedrockPathStyle(catalogProviderID):
+		return sharedllm.NormalizeBedrockModel(modelID)
+	case catalog.IsVertexPathStyle(catalogProviderID):
+		return sharedllm.NormalizeVertexModel(modelID)
+	default:
+		return modelID
+	}
+}

management/internals/modules/agentnetwork/policyselect_model_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/modules/agentnetwork/policyselect_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/modules/agentnetwork/pricing/defaults.go

@@ -47,17 +47,11 @@ var supplementalDefaults = map[string]map[string]Entry{
 		"gpt-5-nano": {InputPer1k: 0.00005, OutputPer1k: 0.0004, CachedInputPer1k: 0.000005},
 	},
 	"anthropic": {
-		// claude-opus-5 is not yet in the catalog lineup but gateway /
-		// grandfathered traffic uses it; priced so it isn't skipped.
-		"claude-opus-5": {InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625},
 		// "kimi-k3[1m]" is the 1M-context alias some Claude Code guides
 		// configure against Moonshot's Anthropic-compatible endpoint;
 		// priced identically to kimi-k3 so those requests aren't skipped.
 		"kimi-k3[1m]": {InputPer1k: 0.003, OutputPer1k: 0.015, CacheReadPer1k: 0.0003},
 	},
-	"bedrock": {
-		"anthropic.claude-opus-5": {InputPer1k: 0.005, OutputPer1k: 0.025, CacheReadPer1k: 0.0005, CacheCreationPer1k: 0.00625},
-	},
 }
 
 var (

management/internals/modules/agentnetwork/pricing/defaults_llm_pricing.example.yaml

@@ -82,6 +82,11 @@ anthropic:
     output_per_1k: 0.015
     cache_read_per_1k: 0.0003
     cache_creation_per_1k: 0.00375
+  claude-sonnet-5:
+    input_per_1k: 0.003
+    output_per_1k: 0.015
+    cache_read_per_1k: 0.0003
+    cache_creation_per_1k: 0.00375
   kimi-k3:
     input_per_1k: 0.003
     output_per_1k: 0.015
@@ -145,6 +150,11 @@ bedrock:
     output_per_1k: 0.015
     cache_read_per_1k: 0.0003
     cache_creation_per_1k: 0.00375
+  anthropic.claude-sonnet-5:
+    input_per_1k: 0.003
+    output_per_1k: 0.015
+    cache_read_per_1k: 0.0003
+    cache_creation_per_1k: 0.00375
   meta.llama3-3-70b-instruct:
     input_per_1k: 0.00072
     output_per_1k: 0.00072

management/internals/modules/agentnetwork/pricing/defaults_test.go

@@ -116,11 +116,13 @@ func TestDefaultTable_PinnedRates(t *testing.T) {
 	assert.InDelta(t, 0.010, fable.InputPer1k, 1e-9, "claude-fable-5 input")
 	assert.InDelta(t, 0.0125, fable.CacheCreationPer1k, 1e-9, "claude-fable-5 cache creation")
 
-	// Supplementals present on their surfaces.
+	// Every id below must stay priced whichever source provides it: the
+	// catalog lineup for the current Claude 5 family, supplementalDefaults
+	// for the ids the dashboard deliberately doesn't offer.
 	for surface, ids := range map[string][]string{
 		"openai":    {"gpt-5", "gpt-5-mini", "gpt-5-nano"},
-		"anthropic": {"claude-opus-5", "kimi-k3[1m]", "kimi-k3"},
-		"bedrock":   {"anthropic.claude-opus-5"},
+		"anthropic": {"claude-opus-5", "claude-sonnet-5", "kimi-k3[1m]", "kimi-k3"},
+		"bedrock":   {"anthropic.claude-opus-5", "anthropic.claude-sonnet-5"},
 	} {
 		for _, id := range ids {
 			_, ok := table[surface][id]

management/internals/modules/agentnetwork/provider_bootstrap_test.go

@@ -1,134 +0,0 @@
-package agentnetwork
-
-import (
-	"context"
-	"runtime"
-	"testing"
-
-	"github.com/golang/mock/gomock"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-
-	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
-	"github.com/netbirdio/netbird/management/server/account"
-	"github.com/netbirdio/netbird/management/server/permissions"
-	"github.com/netbirdio/netbird/management/server/permissions/modules"
-	"github.com/netbirdio/netbird/management/server/permissions/operations"
-	"github.com/netbirdio/netbird/management/server/store"
-	nbtypes "github.com/netbirdio/netbird/management/server/types"
-	"github.com/netbirdio/netbird/shared/management/status"
-)
-
-// bootstrapFixture wires a real sqlite store to a gomock permissions manager
-// so tests can grant the provider permission while denying (or never
-// expecting) the settings one.
-type bootstrapFixture struct {
-	manager Manager
-	store   store.Store
-	perms   *permissions.MockManager
-}
-
-func newBootstrapFixture(t *testing.T) *bootstrapFixture {
-	t.Helper()
-	if runtime.GOOS == "windows" {
-		t.Skip("sqlite store not properly supported on Windows yet")
-	}
-	t.Setenv("NETBIRD_STORE_ENGINE", string(nbtypes.SqliteStoreEngine))
-
-	st, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "", t.TempDir())
-	require.NoError(t, err, "test store setup must succeed")
-	t.Cleanup(cleanUp)
-
-	ctrl := gomock.NewController(t)
-	perms := permissions.NewMockManager(ctrl)
-
-	accounts := account.NewMockManager(ctrl)
-	accounts.EXPECT().StoreEvent(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
-	accounts.EXPECT().UpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
-	accounts.EXPECT().BufferUpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
-
-	return &bootstrapFixture{
-		manager: NewManager(st, perms, accounts, nil),
-		store:   st,
-		perms:   perms,
-	}
-}
-
-func (f *bootstrapFixture) expectPermission(accountID, userID string, module modules.Module, op operations.Operation, allowed bool) {
-	f.perms.EXPECT().
-		ValidateUserPermissions(gomock.Any(), accountID, userID, module, op).
-		Return(allowed, context.Background(), nil)
-}
-
-func newBootstrapProvider(accountID string) *types.Provider {
-	p := types.NewProvider(accountID)
-	p.Name = "openai"
-	p.UpstreamURL = "https://api.openai.com"
-	p.APIKey = "sk-test"
-	p.Enabled = true
-	return p
-}
-
-// TestCreateProviderBootstrapRequiresSettingsPermission pins the gate on the
-// one-time settings bootstrap: creating the first provider with a
-// bootstrap_cluster pins the account's cluster and subdomain, which is a
-// settings write and must not ride on the providers permission alone.
-func TestCreateProviderBootstrapRequiresSettingsPermission(t *testing.T) {
-	ctx := context.Background()
-
-	t.Run("denied without settings permission", func(t *testing.T) {
-		f := newBootstrapFixture(t)
-		f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
-		f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, false)
-
-		_, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "cluster1.example.com")
-		require.Error(t, err, "bootstrap without settings permission must fail")
-		var sErr *status.Error
-		require.ErrorAs(t, err, &sErr)
-		assert.Equal(t, status.PermissionDenied, sErr.Type(), "denial should surface as permission denied")
-
-		providers, err := f.store.GetAccountAgentNetworkProviders(ctx, store.LockingStrengthNone, "account1")
-		require.NoError(t, err)
-		assert.Empty(t, providers, "provider must not be persisted when bootstrap is denied")
-		_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
-		assert.Error(t, err, "settings row must not be created when bootstrap is denied")
-	})
-
-	t.Run("allowed with settings permission", func(t *testing.T) {
-		f := newBootstrapFixture(t)
-		f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
-		f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
-
-		created, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "cluster1.example.com")
-		require.NoError(t, err, "bootstrap with both permissions must succeed")
-		require.NotNil(t, created)
-
-		settings, err := f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
-		require.NoError(t, err, "bootstrap must create the settings row")
-		assert.Equal(t, "cluster1.example.com", settings.Cluster, "settings should pin the bootstrap cluster")
-	})
-
-	t.Run("existing settings need no settings permission", func(t *testing.T) {
-		f := newBootstrapFixture(t)
-		require.NoError(t, f.store.SaveAgentNetworkSettings(ctx, &types.Settings{
-			AccountID: "account1",
-			Cluster:   "cluster1.example.com",
-			Subdomain: "existing",
-		}), "pre-existing settings row setup must succeed")
-
-		// Only the providers permission may be consulted: gomock fails the
-		// test on any unexpected settings-permission call.
-		f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
-
-		_, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "cluster1.example.com")
-		require.NoError(t, err, "create with existing settings must not require the settings permission")
-	})
-
-	t.Run("no bootstrap cluster needs no settings permission", func(t *testing.T) {
-		f := newBootstrapFixture(t)
-		f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
-
-		_, err := f.manager.CreateProvider(ctx, "user1", newBootstrapProvider("account1"), "")
-		require.NoError(t, err, "create without bootstrap must not require the settings permission")
-	})
-}

management/internals/modules/agentnetwork/reconcile.go

@@ -10,6 +10,17 @@ import (
 	"github.com/netbirdio/netbird/shared/management/proto"
 )
 
+// syntheticMapping pairs a synthesised proxy mapping with the address of the
+// proxy that serves it. The cluster is recorded rather than derived from the
+// mapping's domain: ProxyMapping does not carry it, and the previous derivation
+// -- everything after the first DNS label -- is wrong whenever the service's
+// domain is not one label under its proxy's address, which silently addressed
+// updates to a cluster no proxy declares.
+type syntheticMapping struct {
+	mapping *proto.ProxyMapping
+	cluster string
+}
+
 // reconcile recomputes the synthesised reverse-proxy services for an
 // account, diffs them against the previously-synthesised set in the
 // in-memory cache, and emits Create / Update / Delete proxy mappings
@@ -45,18 +56,21 @@ func (m *managerImpl) reconcile(ctx context.Context, accountID string) {
 	}
 
 	oidcCfg := m.proxyController.GetOIDCValidationConfig()
-	current := make(map[string]*proto.ProxyMapping, len(services))
+	current := make(map[string]syntheticMapping, len(services))
 	for _, svc := range services {
 		if svc == nil || svc.ID == "" {
 			continue
 		}
-		current[svc.ID] = svc.ToProtoMapping(rpservice.Update, "", oidcCfg)
+		current[svc.ID] = syntheticMapping{
+			mapping: svc.ToProtoMapping(rpservice.Update, "", oidcCfg),
+			cluster: svc.ProxyCluster,
+		}
 	}
 
 	m.reconcileMu.Lock()
 	previous := m.reconcileCache[accountID]
 	if previous == nil {
-		previous = make(map[string]*proto.ProxyMapping)
+		previous = make(map[string]syntheticMapping)
 	}
 
 	creates, updates, deletes := diffMappings(previous, current)
@@ -67,34 +81,36 @@ func (m *managerImpl) reconcile(ctx context.Context, accountID string) {
 	}
 	m.reconcileMu.Unlock()
 
-	for _, mapping := range creates {
-		mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_CREATED
-		m.proxyController.SendServiceUpdateToCluster(ctx, accountID, mapping, clusterFromMapping(mapping))
+	for _, entry := range creates {
+		entry.mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_CREATED
+		m.proxyController.SendServiceUpdateToCluster(ctx, accountID, entry.mapping, entry.cluster)
 	}
-	for _, mapping := range updates {
-		mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_MODIFIED
-		m.proxyController.SendServiceUpdateToCluster(ctx, accountID, mapping, clusterFromMapping(mapping))
+	for _, entry := range updates {
+		entry.mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_MODIFIED
+		m.proxyController.SendServiceUpdateToCluster(ctx, accountID, entry.mapping, entry.cluster)
 	}
-	for _, mapping := range deletes {
-		mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_REMOVED
-		m.proxyController.SendServiceUpdateToCluster(ctx, accountID, mapping, clusterFromMapping(mapping))
+	for _, entry := range deletes {
+		entry.mapping.Type = proto.ProxyMappingUpdateType_UPDATE_TYPE_REMOVED
+		m.proxyController.SendServiceUpdateToCluster(ctx, accountID, entry.mapping, entry.cluster)
 	}
 }
 
-// diffMappings classifies the previous→current transition for a
-// single account into Create / Update / Delete sets.
+// diffMappings classifies the previous→current transition for a single
+// account into Create / Update / Delete sets.
 //
-// Cluster moves (current.cluster != previous.cluster) are surfaced as
-// a Delete on the old cluster + Create on the new — handled by
-// emitting both a delete (on previous mapping) and a create (on the
-// current mapping) for that service ID.
-func diffMappings(previous, current map[string]*proto.ProxyMapping) (creates, updates, deletes []*proto.ProxyMapping) {
+// A change of serving proxy for the same service ID is surfaced as a Delete
+// addressed to the old proxy plus a Create addressed to the new one, so the
+// mapping actually moves. Comparing the recorded cluster is what makes that
+// detectable: with a placement-free endpoint the mapping's domain is identical
+// before and after the move, so nothing about the mapping itself reveals it.
+func diffMappings(previous, current map[string]syntheticMapping) (creates, updates, deletes []syntheticMapping) {
 	for id, cur := range current {
 		prev, existed := previous[id]
 		switch {
 		case !existed:
 			creates = append(creates, cur)
-		case prev.GetDomain() == "" || cur.GetAccountId() == prev.GetAccountId() && currentClusterChanged(prev, cur):
+		case prev.mapping.GetDomain() == "" ||
+			cur.mapping.GetAccountId() == prev.mapping.GetAccountId() && prev.cluster != cur.cluster:
 			deletes = append(deletes, prev)
 			creates = append(creates, cur)
 		default:
@@ -108,24 +124,3 @@ func diffMappings(previous, current map[string]*proto.ProxyMapping) (creates, up
 	}
 	return creates, updates, deletes
 }
-
-func currentClusterChanged(prev, cur *proto.ProxyMapping) bool {
-	return clusterFromMapping(prev) != clusterFromMapping(cur)
-}
-
-// clusterFromMapping returns the cluster the mapping should be sent
-// to. ProxyMapping doesn't carry the cluster directly, so we rely on
-// the synthesised service's domain (`<slug>.<cluster>`) and split on
-// the first '.'.
-func clusterFromMapping(m *proto.ProxyMapping) string {
-	if m == nil {
-		return ""
-	}
-	domain := m.GetDomain()
-	for i := 0; i < len(domain); i++ {
-		if domain[i] == '.' {
-			return domain[i+1:]
-		}
-	}
-	return ""
-}

management/internals/modules/agentnetwork/reconcile_test.go

@@ -4,7 +4,7 @@ import (
 	"context"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
@@ -21,7 +21,7 @@ func newReconcileMgr(t *testing.T, ctrl *gomock.Controller) (*managerImpl, *stor
 	return &managerImpl{
 		store:           mockStore,
 		proxyController: mockProxy,
-		reconcileCache:  make(map[string]map[string]*proto.ProxyMapping),
+		reconcileCache:  make(map[string]map[string]syntheticMapping),
 	}, mockStore, mockProxy
 }
 
@@ -52,9 +52,9 @@ func newReconcileTestPolicy(providerID, sourceGroupID string) *types.Policy {
 
 func newReconcileTestSettings() *types.Settings {
 	return &types.Settings{
-		AccountID: "acct-1",
-		Cluster:   "eu.proxy.netbird.io",
-		Subdomain: "violet",
+		AccountID:    "acct-1",
+		Domain:       "violet.eu.proxy.netbird.io",
+		ProxyAddress: "eu.proxy.netbird.io",
 	}
 }
 
@@ -196,7 +196,7 @@ func TestReconcile_PolicyRemoved_EmitsDelete(t *testing.T) {
 func TestReconcile_NilProxyController_NoOp(t *testing.T) {
 	ctx := context.Background()
 	mgr := &managerImpl{
-		reconcileCache: make(map[string]map[string]*proto.ProxyMapping),
+		reconcileCache: make(map[string]map[string]syntheticMapping),
 	}
 	// Must not panic; must not query the store.
 	mgr.reconcile(ctx, "acct-1")
@@ -212,21 +212,78 @@ func TestReconcile_EmptyAccountID_NoOp(t *testing.T) {
 	mgr.reconcile(ctx, "")
 }
 
-func TestClusterFromMapping(t *testing.T) {
-	tests := []struct {
-		name   string
-		domain string
-		want   string
-	}{
-		{"simple", "openai.eu.proxy.netbird.io", "eu.proxy.netbird.io"},
-		{"deeply nested", "a.b.c.d", "b.c.d"},
-		{"no dot", "openai", ""},
-		{"empty", "", ""},
+// TestDiffMappings_ServingProxyChange — when the proxy serving an account
+// changes, the same service ID must be deleted on the old proxy and created on
+// the new one. The cluster cannot be recovered from the mapping's domain: with a
+// placement-free endpoint the domain does not change at all when the serving
+// proxy does, so a domain-derived cluster sees no change and emits a plain
+// update, addressed to a proxy that does not exist.
+func TestDiffMappings_ServingProxyChange(t *testing.T) {
+	previous := map[string]syntheticMapping{
+		"svc-1": {
+			mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "brave-otter.gateway.example.com"},
+			cluster: "proxy.example.com",
+		},
 	}
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			got := clusterFromMapping(&proto.ProxyMapping{Domain: tt.domain})
-			assert.Equal(t, tt.want, got)
-		})
+	current := map[string]syntheticMapping{
+		"svc-1": {
+			mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "brave-otter.gateway.example.com"},
+			cluster: "brave-otter.gateway.example.com",
+		},
+	}
+
+	creates, updates, deletes := diffMappings(previous, current)
+
+	if assert.Len(t, deletes, 1, "the old proxy must be told to drop the mapping") {
+		assert.Equal(t, "proxy.example.com", deletes[0].cluster)
+	}
+	if assert.Len(t, creates, 1, "the new proxy must be told to add it") {
+		assert.Equal(t, "brave-otter.gateway.example.com", creates[0].cluster)
+	}
+	assert.Empty(t, updates, "a serving-proxy move is a delete plus a create, not an update")
+}
+
+// TestDiffMappings_UnchangedClusterIsAnUpdate keeps the ordinary path: same
+// service, same proxy, changed contents.
+func TestDiffMappings_UnchangedClusterIsAnUpdate(t *testing.T) {
+	previous := map[string]syntheticMapping{
+		"svc-1": {
+			mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "otter.proxy.example.com"},
+			cluster: "proxy.example.com",
+		},
+	}
+	current := map[string]syntheticMapping{
+		"svc-1": {
+			mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "otter.proxy.example.com"},
+			cluster: "proxy.example.com",
+		},
+	}
+
+	creates, updates, deletes := diffMappings(previous, current)
+
+	assert.Empty(t, creates)
+	assert.Empty(t, deletes)
+	if assert.Len(t, updates, 1) {
+		assert.Equal(t, "proxy.example.com", updates[0].cluster)
+	}
+}
+
+// TestDiffMappings_RemovedServiceIsDeletedOnItsOwnCluster — a service that has
+// gone away is deleted on the cluster it was last served by, which is recorded
+// rather than re-derived.
+func TestDiffMappings_RemovedServiceIsDeletedOnItsOwnCluster(t *testing.T) {
+	previous := map[string]syntheticMapping{
+		"svc-1": {
+			mapping: &proto.ProxyMapping{Id: "svc-1", AccountId: "acct-1", Domain: "brave-otter.gateway.example.com"},
+			cluster: "brave-otter.gateway.example.com",
+		},
+	}
+
+	creates, updates, deletes := diffMappings(previous, map[string]syntheticMapping{})
+
+	assert.Empty(t, creates)
+	assert.Empty(t, updates)
+	if assert.Len(t, deletes, 1) {
+		assert.Equal(t, "brave-otter.gateway.example.com", deletes[0].cluster)
 	}
 }

management/internals/modules/agentnetwork/settings_bootstrap_test.go

@@ -0,0 +1,225 @@
+package agentnetwork
+
+import (
+	"context"
+	"runtime"
+	"strings"
+	"testing"
+
+	"go.uber.org/mock/gomock"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	"github.com/netbirdio/netbird/management/server/account"
+	"github.com/netbirdio/netbird/management/server/permissions"
+	"github.com/netbirdio/netbird/management/server/permissions/modules"
+	"github.com/netbirdio/netbird/management/server/permissions/operations"
+	"github.com/netbirdio/netbird/management/server/store"
+	nbtypes "github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/management/status"
+)
+
+// bootstrapFixture wires a real sqlite store to a gomock permissions manager
+// so tests can grant or deny the settings permission per case.
+type bootstrapFixture struct {
+	manager Manager
+	store   store.Store
+	perms   *permissions.MockManager
+}
+
+func newBootstrapFixture(t *testing.T) *bootstrapFixture {
+	t.Helper()
+	if runtime.GOOS == "windows" {
+		t.Skip("sqlite store not properly supported on Windows yet")
+	}
+	t.Setenv("NETBIRD_STORE_ENGINE", string(nbtypes.SqliteStoreEngine))
+
+	st, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "", t.TempDir())
+	require.NoError(t, err, "test store setup must succeed")
+	t.Cleanup(cleanUp)
+
+	ctrl := gomock.NewController(t)
+	perms := permissions.NewMockManager(ctrl)
+
+	accounts := account.NewMockManager(ctrl)
+	accounts.EXPECT().StoreEvent(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
+	accounts.EXPECT().UpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
+	accounts.EXPECT().BufferUpdateAccountPeers(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
+
+	return &bootstrapFixture{
+		manager: NewManager(st, perms, accounts, nil),
+		store:   st,
+		perms:   perms,
+	}
+}
+
+func (f *bootstrapFixture) expectPermission(accountID, userID string, module modules.Module, op operations.Operation, allowed bool) {
+	f.perms.EXPECT().
+		ValidateUserPermissions(gomock.Any(), accountID, userID, module, op).
+		Return(allowed, context.Background(), nil)
+}
+
+func (f *bootstrapFixture) createSettings(ctx context.Context, accountID, userID, proxyAddress, endpoint string) (*types.Settings, error) {
+	return f.manager.CreateSettings(ctx, userID, types.DefaultSettings(accountID), proxyAddress, endpoint)
+}
+
+// TestCreateSettingsRequiresPermission pins the gate: bootstrap assigns the
+// account's immutable endpoint, a settings write requiring the settings
+// Create permission — and a denial leaves no row behind.
+func TestCreateSettingsRequiresPermission(t *testing.T) {
+	ctx := context.Background()
+	f := newBootstrapFixture(t)
+	f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, false)
+
+	_, err := f.createSettings(ctx, "account1", "user1", "cluster1.example.com", "")
+	require.Error(t, err, "bootstrap without the settings permission must fail")
+	var sErr *status.Error
+	require.ErrorAs(t, err, &sErr)
+	assert.Equal(t, status.PermissionDenied, sErr.Type(), "denial should surface as permission denied")
+
+	_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
+	assert.Error(t, err, "settings row must not be created when bootstrap is denied")
+}
+
+// TestCreateSettingsLabeled pins the labeled shape: the server allocates an
+// adjective-noun label beneath the proxy address, the pin is not dedicated,
+// and the domain records the full endpoint hostname.
+func TestCreateSettingsLabeled(t *testing.T) {
+	ctx := context.Background()
+	f := newBootstrapFixture(t)
+	f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
+
+	created, err := f.createSettings(ctx, "account1", "user1", "Cluster1.Example.com", "")
+	require.NoError(t, err, "labeled bootstrap must succeed")
+	assert.Equal(t, "cluster1.example.com", created.ProxyAddress, "proxy address must be pinned lowercased")
+	require.True(t, strings.HasSuffix(created.Domain, ".cluster1.example.com"),
+		"domain must hang one label beneath the proxy address: %s", created.Domain)
+	label := strings.TrimSuffix(created.Domain, ".cluster1.example.com")
+	assert.NotContains(t, label, ".", "the allocated label must be a single DNS label: %s", label)
+	assert.False(t, created.Dedicated(), "a labeled pin is not dedicated")
+	assert.Equal(t, created.Domain, created.Endpoint(), "the endpoint is the domain column")
+
+	stored, err := f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
+	require.NoError(t, err, "bootstrap must persist the row")
+	assert.Equal(t, created.Domain, stored.Domain)
+	assert.Equal(t, created.ProxyAddress, stored.ProxyAddress)
+}
+
+// TestCreateSettingsSelfAddressed pins the dedicated shape: the endpoint is
+// claimed verbatim (normalized), Domain == ProxyAddress, and the claim
+// succeeds with no proxy declaring the address yet (address-first).
+func TestCreateSettingsSelfAddressed(t *testing.T) {
+	ctx := context.Background()
+	f := newBootstrapFixture(t)
+	f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
+
+	created, err := f.createSettings(ctx, "account1", "user1", "", "Brave-Otter.GW.Example.com")
+	require.NoError(t, err, "self-addressed bootstrap must succeed")
+	assert.Equal(t, "brave-otter.gw.example.com", created.Domain, "endpoint must be claimed lowercased")
+	assert.Equal(t, created.Domain, created.ProxyAddress, "self-addressed: proxy address is the endpoint")
+	assert.True(t, created.Dedicated(), "a self-addressed pin is dedicated")
+}
+
+// TestCreateSettingsIdentityFieldValidation pins the request contract: exactly
+// one of proxyAddress and endpoint, and both must be well-formed hostnames.
+func TestCreateSettingsIdentityFieldValidation(t *testing.T) {
+	ctx := context.Background()
+
+	cases := map[string]struct {
+		proxyAddress string
+		endpoint     string
+	}{
+		"neither":               {"", ""},
+		"both":                  {"cluster1.example.com", "gw.example.com"},
+		"trailing dot endpoint": {"", "gw.example.com."},
+		"leading dot endpoint":  {"", ".gw.example.com"},
+		"whitespace inside":     {"", "g w.example.com"},
+		"empty label in parent": {"eu..example.com", ""},
+		"hyphen-edged label":    {"", "-gw.example.com"},
+	}
+	for name, tc := range cases {
+		t.Run(name, func(t *testing.T) {
+			f := newBootstrapFixture(t)
+			f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
+
+			_, err := f.createSettings(ctx, "account1", "user1", tc.proxyAddress, tc.endpoint)
+			require.Error(t, err, "invalid identity input must be rejected")
+			var sErr *status.Error
+			require.ErrorAs(t, err, &sErr)
+			assert.Equal(t, status.InvalidArgument, sErr.Type(), "rejection must be a validation error")
+
+			_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
+			assert.Error(t, err, "no row may be left behind by a rejected bootstrap")
+		})
+	}
+}
+
+// TestCreateSettingsConflictsOnSecondBootstrap pins that bootstrap is a
+// one-time create per account: a second call is a conflict, whatever shape it
+// asks for, and the original row survives untouched.
+func TestCreateSettingsConflictsOnSecondBootstrap(t *testing.T) {
+	ctx := context.Background()
+	f := newBootstrapFixture(t)
+	f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
+
+	first, err := f.createSettings(ctx, "account1", "user1", "cluster1.example.com", "")
+	require.NoError(t, err)
+
+	f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
+	_, err = f.createSettings(ctx, "account1", "user1", "", "other.example.com")
+	require.Error(t, err, "second bootstrap must fail")
+	var sErr *status.Error
+	require.ErrorAs(t, err, &sErr)
+	assert.Equal(t, status.AlreadyExists, sErr.Type(), "second bootstrap must surface as a conflict")
+
+	stored, err := f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
+	require.NoError(t, err)
+	assert.Equal(t, first.Domain, stored.Domain, "the original endpoint must survive the rejected bootstrap")
+}
+
+// TestCreateSettingsEndpointTaken pins global hostname uniqueness: a hostname
+// held by one account cannot be claimed by another, in either direction —
+// self-addressed onto self-addressed, or self-addressed onto an allocated
+// labeled endpoint.
+func TestCreateSettingsEndpointTaken(t *testing.T) {
+	ctx := context.Background()
+	f := newBootstrapFixture(t)
+
+	f.expectPermission("account1", "user1", modules.AgentNetworkSettings, operations.Create, true)
+	first, err := f.createSettings(ctx, "account1", "user1", "", "gw.example.com")
+	require.NoError(t, err)
+
+	f.expectPermission("account2", "user2", modules.AgentNetworkSettings, operations.Create, true)
+	_, err = f.createSettings(ctx, "account2", "user2", "", "gw.example.com")
+	require.Error(t, err, "a taken hostname must be refused")
+	var sErr *status.Error
+	require.ErrorAs(t, err, &sErr)
+	assert.Equal(t, status.AlreadyExists, sErr.Type(), "the refusal must surface as a conflict")
+
+	f.expectPermission("account3", "user3", modules.AgentNetworkSettings, operations.Create, true)
+	_, err = f.createSettings(ctx, "account3", "user3", "", first.Domain)
+	require.Error(t, err, "claiming another account's endpoint must be refused")
+}
+
+// TestCreateProviderHasNoSettingsSideEffects pins the decoupling: provider
+// create needs only the providers permission (gomock fails the test on any
+// settings-permission call) and never creates a settings row.
+func TestCreateProviderHasNoSettingsSideEffects(t *testing.T) {
+	ctx := context.Background()
+	f := newBootstrapFixture(t)
+	f.expectPermission("account1", "user1", modules.AgentNetworkProviders, operations.Create, true)
+
+	provider := types.NewProvider("account1")
+	provider.Name = "openai"
+	provider.UpstreamURL = "https://api.openai.com"
+	provider.APIKey = "sk-test"
+	provider.Enabled = true
+
+	created, err := f.manager.CreateProvider(ctx, "user1", provider)
+	require.NoError(t, err, "provider create must succeed on the providers permission alone")
+	require.NotNil(t, created)
+
+	_, err = f.store.GetAgentNetworkSettings(ctx, store.LockingStrengthNone, "account1")
+	assert.Error(t, err, "provider create must not conjure a settings row")
+}

management/internals/modules/agentnetwork/synthesizer.go

@@ -10,6 +10,7 @@ import (
 	"strings"
 
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/catalog"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/modeldiscovery"
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
 	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/sessionkey"
@@ -89,7 +90,7 @@ func SynthesizeServicesForCluster(ctx context.Context, s store.Store, clusterAdd
 		return nil, nil
 	}
 
-	settingsRows, err := s.GetAgentNetworkSettingsByCluster(ctx, store.LockingStrengthNone, clusterAddr)
+	settingsRows, err := s.GetAgentNetworkSettingsByProxyAddress(ctx, store.LockingStrengthNone, clusterAddr)
 	if err != nil {
 		return nil, fmt.Errorf("list agent network settings on cluster: %w", err)
 	}
@@ -116,53 +117,41 @@ func SynthesizeServicesForCluster(ctx context.Context, s store.Store, clusterAdd
 }
 
 // SynthesizeServiceForDomain resolves a single agent-network service by its
-// public endpoint domain. It lists the (few) settings rows on the domain's
-// cluster, matches the one whose endpoint equals the domain, and synthesises
-// only that account — avoiding full per-account synthesis for every tenant on
-// the cluster, which is what auth/session paths previously paid. Returns nil
-// (no error) when no account owns the domain.
+// public endpoint domain — a point query on the settings domain unique index,
+// then synthesis of just that account. Returns nil (no error) when no account
+// owns the domain.
 func SynthesizeServiceForDomain(ctx context.Context, s store.Store, domain string) (*rpservice.Service, error) {
-	domain = strings.TrimSpace(domain)
-	cluster := clusterFromDomain(domain)
-	if domain != "" && cluster != "" {
-		settingsRows, err := s.GetAgentNetworkSettingsByCluster(ctx, store.LockingStrengthNone, cluster)
-		if err != nil {
-			return nil, fmt.Errorf("list agent network settings on cluster: %w", err)
+	domain = strings.ToLower(strings.TrimSpace(domain))
+	if domain == "" {
+		return nil, nil //nolint:nilnil // optional lookup: no account owns the domain
+	}
+
+	settings, err := s.GetAgentNetworkSettingsByDomain(ctx, store.LockingStrengthNone, domain)
+	if err != nil {
+		if isNotFound(err) {
+			return nil, nil //nolint:nilnil // optional lookup: no account owns the domain
 		}
-		for _, settings := range settingsRows {
-			if settings == nil || settings.Endpoint() != domain {
-				continue
-			}
-			services, serr := SynthesizeServices(ctx, s, settings.AccountID)
-			if serr != nil {
-				return nil, serr
-			}
-			for _, svc := range services {
-				if svc != nil && svc.Domain == domain {
-					return svc, nil
-				}
-			}
-			break
+		return nil, fmt.Errorf("get agent network settings by domain: %w", err)
+	}
+
+	services, err := SynthesizeServices(ctx, s, settings.AccountID)
+	if err != nil {
+		return nil, err
+	}
+	for _, svc := range services {
+		if svc != nil && svc.Domain == domain {
+			return svc, nil
 		}
 	}
 	return nil, nil //nolint:nilnil // optional lookup: no account owns the domain
 }
 
-// clusterFromDomain returns the cluster portion of an endpoint domain (every
-// label after the first).
-func clusterFromDomain(domain string) string {
-	if i := strings.IndexByte(domain, '.'); i >= 0 {
-		return domain[i+1:]
-	}
-	return ""
-}
-
 // SynthesizeServices builds the in-memory reverse-proxy service that
 // fronts the account's agent-network gateway. Returns nil when the
 // account has no settings row, no enabled providers, or no enabled
 // policies — in any of those cases there's nothing useful to expose.
 //
-// One service per (account, settings.Cluster) is emitted. The router
+// One service per (account, settings.ProxyAddress) is emitted. The router
 // middleware encodes a denormalised model→provider routing table
 // (auth headers + decrypted API keys baked in); the policy_check
 // middleware encodes per-provider authorised group IDs derived from
@@ -175,7 +164,7 @@ func SynthesizeServices(ctx context.Context, s store.Store, accountID string) ([
 	if err != nil {
 		return nil, err
 	}
-	if !ok || strings.TrimSpace(settings.Cluster) == "" {
+	if !ok || strings.TrimSpace(settings.ProxyAddress) == "" {
 		return nil, nil
 	}
 
@@ -223,7 +212,19 @@ func SynthesizeServices(ctx context.Context, s store.Store, accountID string) ([
 
 	groupIndex := indexProviderGroups(enabledPolicies)
 
-	routerCfgJSON, err := buildRouterConfigJSON(enabledProviders, groupIndex)
+	// The proxy guardrail is a per-provider fail-closed backstop; the
+	// authoritative per-policy/group decision is management's
+	// SelectPolicyForRequest. A provider lands in that map only when every
+	// authorising policy restricts models.
+	providerAllowlists := buildProviderAllowlists(enabledPolicies, guardrailsByID)
+
+	// Discovery gets the finer view: per policy rather than flattened per
+	// provider, so a listing can be bounded to what the calling groups may
+	// actually use instead of the union across everyone who reaches the
+	// provider.
+	modelPolicies := buildModelPolicies(enabledPolicies, guardrailsByID)
+
+	routerCfgJSON, err := buildRouterConfigJSON(enabledProviders, groupIndex, modelPolicies)
 	if err != nil {
 		return nil, err
 	}
@@ -240,11 +241,6 @@ func SynthesizeServices(ctx context.Context, s store.Store, accountID string) ([
 
 	mergedGuardrails := mergeGuardrails(enabledPolicies, guardrailsByID)
 	applyAccountCollectionControls(&mergedGuardrails, settings)
-	// The proxy guardrail is a per-provider fail-closed backstop; the
-	// authoritative per-policy/group decision is management's
-	// SelectPolicyForRequest. A provider lands in this map only when every
-	// authorising policy restricts models.
-	providerAllowlists := buildProviderAllowlists(enabledPolicies, guardrailsByID)
 	guardrailJSON, err := marshalGuardrailConfig(providerAllowlists, mergedGuardrails.PromptCapture)
 	if err != nil {
 		return nil, err
@@ -363,6 +359,11 @@ type routerProviderRoute struct {
 	AuthHeaderName  string   `json:"auth_header_name"`
 	AuthHeaderValue string   `json:"auth_header_value"`
 	AllowedGroupIDs []string `json:"allowed_group_ids,omitempty"`
+	// ModelPolicies is one entry per enabled policy authorising this provider,
+	// carrying that policy's source groups and the models it permits. The
+	// router bounds a model listing with it, so a provider two groups reach
+	// under different allowlists offers each only its own.
+	ModelPolicies []routerModelPolicy `json:"model_policies,omitempty"`
 	// Vertex marks a Google Vertex AI provider, whose requests carry the
 	// model in the URL path. The router selects it by path, bypassing the
 	// model/vendor table.
@@ -380,6 +381,9 @@ type routerProviderRoute struct {
 	// proxy dials this provider's upstream. For self-hosted / internal gateways
 	// behind a private or self-signed certificate.
 	SkipTLSVerify bool `json:"skip_tls_verify,omitempty"`
+	// DiscoveryHost, when set, is the host serving this provider's model
+	// listing, for a vendor that does not serve it from the inference host.
+	DiscoveryHost string `json:"discovery_host,omitempty"`
 }
 
 // indexProviderGroups walks the enabled policies and returns, per
@@ -434,7 +438,7 @@ func indexProviderGroups(policies []*types.Policy) map[string][]string {
 // path-prefix tiebreak. Providers no enabled policy authorises
 // (orphans) are intentionally OMITTED so the router never observes a
 // route with an empty ACL.
-func buildRouterConfigJSON(providers []*types.Provider, groupIndex map[string][]string) ([]byte, error) {
+func buildRouterConfigJSON(providers []*types.Provider, groupIndex map[string][]string, modelPolicies map[string][]routerModelPolicy) ([]byte, error) {
 	cfg := routerConfig{Providers: make([]routerProviderRoute, 0, len(providers))}
 	for _, p := range providers {
 		groups, hasPolicy := groupIndex[p.ID]
@@ -447,6 +451,9 @@ func buildRouterConfigJSON(providers []*types.Provider, groupIndex map[string][]
 		if err != nil {
 			return nil, fmt.Errorf("router config for provider %s: %w", p.ID, err)
 		}
+		// Lookup rather than assume: an unknown provider id yields the zero
+		// entry, which declares no discovery and so contributes nothing.
+		catalogEntry, _ := catalog.Lookup(p.ProviderID)
 		headerName, headerValue, gcpSAKeyB64, err := providerAuthHeader(p)
 		if err != nil {
 			return nil, err
@@ -461,10 +468,12 @@ func buildRouterConfigJSON(providers []*types.Provider, groupIndex map[string][]
 			AuthHeaderName:          headerName,
 			AuthHeaderValue:         headerValue,
 			AllowedGroupIDs:         groups,
+			ModelPolicies:           modelPolicies[p.ID],
 			Vertex:                  catalog.IsVertexPathStyle(p.ProviderID),
 			Bedrock:                 catalog.IsBedrockPathStyle(p.ProviderID),
 			GCPServiceAccountKeyB64: gcpSAKeyB64,
 			SkipTLSVerify:           p.SkipTLSVerification,
+			DiscoveryHost:           discoveryHost(catalogEntry, p.UpstreamURL),
 		})
 	}
 	out, err := json.Marshal(cfg)
@@ -474,6 +483,33 @@ func buildRouterConfigJSON(providers []*types.Provider, groupIndex map[string][]
 	return out, nil
 }
 
+// discoveryHost returns the host serving this provider's model listing when it
+// differs from the inference host, and empty when the two are the same — which
+// is true of every vendor but Bedrock, whose ListInferenceProfiles is a control
+// plane operation on bedrock.<region> while InvokeModel must go to
+// bedrock-runtime.<region>. One provider record therefore needs two hosts.
+//
+// The catalog declares the listing host; the region is recovered from the
+// upstream the operator configured, since a provider record carries no region
+// field. An upstream matching no catalog template yields empty rather than a
+// guess: a proxied or self-hosted Bedrock endpoint may serve both from one
+// place, and inventing a host would send the credential somewhere the operator
+// never configured.
+func discoveryHost(entry catalog.Provider, upstreamURL string) string {
+	if entry.Discovery == nil || entry.Discovery.Host == "" {
+		return ""
+	}
+	host := entry.Discovery.Host
+	if !strings.Contains(host, catalog.RegionPlaceholder) {
+		return host
+	}
+	region := modeldiscovery.RegionFromUpstream(entry, upstreamURL)
+	if region == "" {
+		return ""
+	}
+	return strings.ReplaceAll(host, catalog.RegionPlaceholder, region)
+}
+
 // providerVendor returns the parser surface ("openai", "anthropic", …)
 // the provider speaks, sourced from its catalog entry's ParserID. The
 // router uses it to keep a request the parser tagged with a vendor on a
@@ -934,7 +970,7 @@ func buildAccountService(
 	middlewares []rpservice.MiddlewareConfig,
 	sessionPriv, sessionPub string,
 ) *rpservice.Service {
-	cluster := settings.Cluster
+	cluster := settings.ProxyAddress
 	domain := settings.Endpoint()
 	serviceID := SynthesizedServiceIDPrefix + accountID
 
@@ -1110,3 +1146,46 @@ func mergeGuardrail(g *types.Guardrail, merged *MergedGuardrails) {
 		}
 	}
 }
+
+// routerModelPolicy mirrors the router's ModelPolicyRule: one authorising
+// policy's source groups plus the models it permits. Models is nil for a
+// policy that sets no model allowlist, which lifts the restriction for the
+// groups it binds — so nil and empty must survive the round trip distinctly.
+type routerModelPolicy struct {
+	GroupIDs []string `json:"group_ids"`
+	Models   []string `json:"models"`
+}
+
+// buildModelPolicies indexes, per provider, one rule for each enabled policy
+// authorising it: the policy's source groups and the models its guardrail
+// permits.
+//
+// This is deliberately finer than buildProviderAllowlists, which flattens the
+// same inputs into one list per provider for the proxy's fail-closed guardrail.
+// A flattened list cannot answer "what may THIS caller see", so a provider two
+// teams reach under different allowlists would offer each team the other's
+// models — a picker full of entries the next request refuses. Keeping the
+// source groups alongside the models lets the router answer it at request time,
+// where it knows the caller's groups.
+func buildModelPolicies(policies []*types.Policy, byID map[string]*types.Guardrail) map[string][]routerModelPolicy {
+	out := make(map[string][]routerModelPolicy)
+	for _, p := range policies {
+		if p == nil || len(p.SourceGroups) == 0 {
+			continue
+		}
+		restricted, models := policyModelAllowlist(p, byID)
+		rule := routerModelPolicy{GroupIDs: append([]string(nil), p.SourceGroups...)}
+		if restricted {
+			// Never nil when restricted: an allowlist permitting nothing must
+			// stay distinguishable from no allowlist at all.
+			rule.Models = append([]string{}, models...)
+		}
+		for _, providerID := range p.DestinationProviderIDs {
+			if providerID == "" {
+				continue
+			}
+			out[providerID] = append(out[providerID], rule)
+		}
+	}
+	return out
+}

management/internals/modules/agentnetwork/synthesizer_pricing_test.go

@@ -103,3 +103,37 @@ func TestBuildCostMeterConfig_OrphanAndGatewayProviders(t *testing.T) {
 	assert.NotContains(t, cfg.Pricing.Providers, "prov-litellm", "empty-models gateway needs no per-record entry")
 	assert.NotEmpty(t, cfg.Pricing.Defaults["openai"], "defaults still ship so the gateway's catalog-model traffic is priced")
 }
+
+// TestBuildCostMeterConfig_BedrockGeographyOutsideTheOriginalFour is the
+// accounting half of the geography bug. The docs tell operators to register a
+// Bedrock id exactly as AWS issues it, region prefix included, and the cost
+// meter keys its table by the normalized form. While the geography was matched
+// against a list of four, a profile issued anywhere else kept its prefix,
+// missed the catalog entry it was meant to inherit from, and billed with a
+// zero entry underneath the operator's own rates — so every cache bucket
+// metered free and a model priced only by catalog defaults metered at nothing
+// at all.
+func TestBuildCostMeterConfig_BedrockGeographyOutsideTheOriginalFour(t *testing.T) {
+	for _, geo := range []string{"jp", "au", "ca", "sa", "us-gov"} {
+		t.Run(geo, func(t *testing.T) {
+			bedrock := &types.Provider{
+				ID:         "prov-bedrock",
+				ProviderID: "bedrock_api",
+				Enabled:    true,
+				Models: []types.ProviderModel{
+					{ID: geo + ".anthropic.claude-sonnet-5-20260514-v1:0", InputPer1k: 0.003, OutputPer1k: 0.015},
+				},
+			}
+			raw, err := buildCostMeterConfigJSON([]*types.Provider{bedrock}, map[string][]string{"prov-bedrock": {"grp"}})
+			require.NoError(t, err)
+			cfg := decodeCostMeterConfig(t, raw)
+
+			e, ok := cfg.Pricing.Providers["prov-bedrock"]["anthropic.claude-sonnet-5"]
+			require.True(t, ok, "a %s profile must key by the same normalized id the parser emits", geo)
+			assert.InDelta(t, 0.0003, e.CacheReadPer1k, 1e-9,
+				"cache read must be inherited from the bedrock default entry, not left at zero")
+			assert.InDelta(t, 0.00375, e.CacheCreationPer1k, 1e-9,
+				"cache creation must be inherited from the bedrock default entry, not left at zero")
+		})
+	}
+}

management/internals/modules/agentnetwork/synthesizer_provider_allowlist_test.go

@@ -4,6 +4,7 @@ import (
 	"testing"
 
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
 )
@@ -93,3 +94,75 @@ func TestBuildProviderAllowlists(t *testing.T) {
 			"an enabled-but-empty allowlist is restricted with an empty set, not unrestricted")
 	})
 }
+
+// policyForGroups builds an enabled policy binding the given source groups to
+// the given providers under an optional guardrail.
+func policyForGroups(id string, groups []string, guardrailIDs []string, providerIDs ...string) *types.Policy {
+	return &types.Policy{
+		ID:                     id,
+		Enabled:                true,
+		SourceGroups:           groups,
+		DestinationProviderIDs: providerIDs,
+		GuardrailIDs:           guardrailIDs,
+	}
+}
+
+// TestBuildModelPolicies covers the finer index discovery needs. Where
+// buildProviderAllowlists flattens every authorising policy into one list per
+// provider — enough for a fail-closed backstop, but blind to who is asking —
+// this keeps each policy's source groups beside its models so the router can
+// bound a listing to the calling groups.
+func TestBuildModelPolicies(t *testing.T) {
+	byID := map[string]*types.Guardrail{
+		"g-4o":       allowlistGuardrail("g-4o", "acc-1", "gpt-4o"),
+		"g-opus":     allowlistGuardrail("g-opus", "acc-1", "claude-opus-4"),
+		"g-disabled": {ID: "g-disabled", Checks: types.GuardrailChecks{ModelAllowlist: types.GuardrailModelAllowlist{Enabled: false, Models: []string{"gpt-4o"}}}},
+	}
+
+	t.Run("each policy keeps its own groups and models", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForGroups("p1", []string{"grp-eng"}, []string{"g-4o"}, "prov-x"),
+			policyForGroups("p2", []string{"grp-sales"}, []string{"g-opus"}, "prov-x"),
+		}
+		got := buildModelPolicies(policies, byID)
+		assert.Equal(t, []routerModelPolicy{
+			{GroupIDs: []string{"grp-eng"}, Models: []string{"gpt-4o"}},
+			{GroupIDs: []string{"grp-sales"}, Models: []string{"claude-opus-4"}},
+		}, got["prov-x"],
+			"the two policies must stay separable so neither group is offered the other's models")
+	})
+
+	t.Run("an unrestricted policy carries nil models", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForGroups("p1", []string{"grp-eng"}, []string{"g-4o"}, "prov-x"),
+			policyForGroups("p2", []string{"grp-admin"}, nil, "prov-x"),
+		}
+		got := buildModelPolicies(policies, byID)
+		assert.Nil(t, got["prov-x"][1].Models,
+			"no allowlist must reach the router as nil, which lifts the restriction for its groups")
+	})
+
+	t.Run("a disabled allowlist is not a restriction", func(t *testing.T) {
+		policies := []*types.Policy{policyForGroups("p1", []string{"grp-eng"}, []string{"g-disabled"}, "prov-x")}
+		got := buildModelPolicies(policies, byID)
+		assert.Nil(t, got["prov-x"][0].Models,
+			"a guardrail with the allowlist check off restricts nothing")
+	})
+
+	t.Run("an enabled allowlist with no models permits nothing", func(t *testing.T) {
+		byIDEmpty := map[string]*types.Guardrail{
+			"g-empty": {ID: "g-empty", Checks: types.GuardrailChecks{ModelAllowlist: types.GuardrailModelAllowlist{Enabled: true}}},
+		}
+		policies := []*types.Policy{policyForGroups("p1", []string{"grp-eng"}, []string{"g-empty"}, "prov-x")}
+		got := buildModelPolicies(policies, byIDEmpty)
+		require.NotNil(t, got["prov-x"][0].Models,
+			"an empty allowlist must not arrive as nil — that would read as unrestricted")
+		assert.Empty(t, got["prov-x"][0].Models)
+	})
+
+	t.Run("a policy binding no groups is skipped", func(t *testing.T) {
+		policies := []*types.Policy{policyForGroups("p1", nil, []string{"g-4o"}, "prov-x")}
+		assert.Empty(t, buildModelPolicies(policies, byID),
+			"a policy with no source groups authorises nobody, so it bounds nobody's listing")
+	})
+}

management/internals/modules/agentnetwork/synthesizer_realstore_test.go

@@ -147,7 +147,7 @@ func TestReconcile_RealStore_PushesPrivateAfterStatusToggle(t *testing.T) {
 		store:           s,
 		accountManager:  noopAccountManager{},
 		proxyController: ctrl,
-		reconcileCache:  make(map[string]map[string]*proto.ProxyMapping),
+		reconcileCache:  make(map[string]map[string]syntheticMapping),
 	}
 
 	m.reconcile(ctx, testAccountID) // initial, provider enabled

management/internals/modules/agentnetwork/synthesizer_test.go

@@ -6,10 +6,11 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/catalog"
 	"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/store"
@@ -19,15 +20,14 @@ import (
 const (
 	testAccountID = "acct-1"
 	testCluster   = "eu.proxy.netbird.io"
-	testSubdomain = "violet"
 	testEndpoint  = "violet.eu.proxy.netbird.io"
 )
 
 func newSynthTestSettings() *types.Settings {
 	return &types.Settings{
-		AccountID: testAccountID,
-		Cluster:   testCluster,
-		Subdomain: testSubdomain,
+		AccountID:    testAccountID,
+		Domain:       testEndpoint,
+		ProxyAddress: testCluster,
 	}
 }
 
@@ -1246,3 +1246,57 @@ func TestSynthesizeServices_EmptyAPIKey_FailsClosed(t *testing.T) {
 	require.Error(t, err, "synthesis must refuse a provider with no api key")
 	assert.Contains(t, err.Error(), "no api key", "error must surface the missing credential")
 }
+
+// TestDiscoveryHost pins which providers get a separate listing host. Getting
+// this wrong in either direction is costly: a missing host leaves Bedrock
+// discovery 404ing at AWS, and a host on the wrong provider would send that
+// provider's listing — and its credential — somewhere the operator never
+// configured.
+func TestDiscoveryHost(t *testing.T) {
+	entry := func(id string) catalog.Provider {
+		p, ok := catalog.Lookup(id)
+		require.True(t, ok, "catalog entry %s must exist", id)
+		return p
+	}
+
+	for _, tc := range []struct {
+		name     string
+		entry    catalog.Provider
+		upstream string
+		want     string
+	}{
+		{
+			// ListInferenceProfiles is a control-plane operation; the runtime
+			// host answers <UnknownOperationException/> for it.
+			name:  "bedrock splits the listing off the runtime host",
+			entry: entry("bedrock_api"), upstream: "https://bedrock-runtime.eu-central-1.amazonaws.com",
+			want: "bedrock.eu-central-1.amazonaws.com",
+		},
+		{
+			name:  "bedrock in another region",
+			entry: entry("bedrock_api"), upstream: "https://bedrock-runtime.us-west-2.amazonaws.com",
+			want: "bedrock.us-west-2.amazonaws.com",
+		},
+		{
+			// A proxied Bedrock endpoint may well serve both from one place,
+			// and there is no region to read back out of it.
+			name:  "proxied bedrock upstream yields no discovery host",
+			entry: entry("bedrock_api"), upstream: "https://bedrock.internal.example.com",
+			want: "",
+		},
+		{
+			name:  "openai serves its listing from the same host",
+			entry: entry("openai_api"), upstream: "https://api.openai.com",
+			want: "",
+		},
+		{
+			name:  "vertex serves its listing from the same host",
+			entry: entry("vertex_ai_api"), upstream: "https://us-east5-aiplatform.googleapis.com",
+			want: "",
+		},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			assert.Equal(t, tc.want, discoveryHost(tc.entry, tc.upstream))
+		})
+	}
+}

management/internals/modules/agentnetwork/types/settings.go

@@ -1,6 +1,7 @@
 package types
 
 import (
+	"fmt"
 	"strings"
 	"time"
 
@@ -12,13 +13,23 @@ import (
 // the long-term aggregate and are retained independently.
 const DefaultAccessLogRetentionDays = 30
 
-// Settings is the per-account agent-network configuration row. One
-// row per account. Cluster + Subdomain are immutable once written and
-// produce the public endpoint agents call (`<subdomain>.<cluster>`).
+// Settings is the per-account agent-network configuration row. One row per
+// account. Domain and ProxyAddress are assigned at bootstrap and immutable
+// thereafter; a persisted row is always fully allocated — there is no "row
+// exists, endpoint pending" state.
 type Settings struct {
 	AccountID string `gorm:"primaryKey"`
-	Cluster   string
-	Subdomain string `gorm:"index:idx_agent_network_settings_cluster_subdomain"`
+
+	// Domain is the gateway endpoint hostname agents call. Globally unique
+	// across accounts. Sized explicitly because MySQL cannot index an
+	// unbounded TEXT column; 255 covers the RFC 1035 253-octet bound.
+	Domain string `gorm:"type:varchar(255);uniqueIndex:idx_agent_network_settings_domain"`
+
+	// ProxyAddress is the declared cluster address of the proxy serving this
+	// account's gateway. Either equal to Domain — a proxy dedicated to this
+	// account, declaring the tenant's own hostname — or Domain's immediate
+	// parent, with the endpoint one label beneath it on a shared cluster.
+	ProxyAddress string `gorm:"type:varchar(255);index:idx_agent_network_settings_proxy_address"`
 
 	// Account-level collection controls sourced by the synthesizer.
 	// EnableLogCollection gates the per-request access-log trail and defaults
@@ -45,9 +56,9 @@ func (Settings) TableName() string { return "agent_network_settings" }
 
 // DefaultSettings returns the settings an account observes before its row is
 // bootstrapped: log collection on with the default retention, everything else
-// off, and no cluster/subdomain assigned yet. Bootstrap persists exactly these
-// values plus the assigned cluster and subdomain, so the pre-bootstrap read
-// and the freshly bootstrapped row agree.
+// off, and no domain or proxy address assigned yet. Bootstrap persists exactly
+// these values plus the assigned domain and proxy address, so the
+// pre-bootstrap read and the freshly bootstrapped row agree.
 func DefaultSettings(accountID string) *Settings {
 	return &Settings{
 		AccountID:              accountID,
@@ -56,14 +67,15 @@ func DefaultSettings(accountID string) *Settings {
 	}
 }
 
-// Endpoint returns the bare hostname agents reach this account at:
-// `<subdomain>.<cluster>`. Empty until both halves are assigned at bootstrap.
-func (s *Settings) Endpoint() string {
-	if s.Cluster == "" || s.Subdomain == "" {
-		return ""
-	}
-	return s.Subdomain + "." + s.Cluster
-}
+// Endpoint returns the bare hostname agents reach this account at — the
+// Domain column. Empty until the row is bootstrapped.
+func (s *Settings) Endpoint() string { return s.Domain }
+
+// Dedicated reports whether the account's gateway is served by a proxy
+// dedicated to it — the self-addressed shape, where the serving proxy declares
+// the endpoint hostname itself. The alternative (labeled) shape has the
+// endpoint one label beneath a shared cluster's address.
+func (s *Settings) Dedicated() bool { return s.Domain != "" && s.Domain == s.ProxyAddress }
 
 // ToAPIResponse renders the settings as the API representation. The
 // timestamps are omitted while zero — a default (not yet bootstrapped) view
@@ -71,9 +83,9 @@ func (s *Settings) Endpoint() string {
 func (s *Settings) ToAPIResponse() *api.AgentNetworkSettings {
 	retention := s.AccessLogRetentionDays
 	resp := &api.AgentNetworkSettings{
-		Cluster:                s.Cluster,
-		Subdomain:              s.Subdomain,
 		Endpoint:               s.Endpoint(),
+		ProxyAddress:           s.ProxyAddress,
+		Dedicated:              s.Dedicated(),
 		EnableLogCollection:    s.EnableLogCollection,
 		EnablePromptCollection: s.EnablePromptCollection,
 		RedactPii:              s.RedactPii,
@@ -90,19 +102,91 @@ func (s *Settings) ToAPIResponse() *api.AgentNetworkSettings {
 	return resp
 }
 
-// FromAPIRequest applies the request onto the receiver. The mutable
-// collection fields are always replaced with the request values. Cluster
-// participates only in bootstrap and the immutability check (see
-// Manager.UpdateSettings); Subdomain is server-assigned and never taken
-// from a request.
+// FromAPIRequest applies the update request onto the receiver: every mutable
+// field is replaced with the request value, and the identity fields (Domain,
+// ProxyAddress) carry the request's echo of the assigned values. The identity
+// fields are never written to the stored row — UpdateSettings compares them
+// against it and rejects the request when they differ, so PUT keeps the
+// house convention of requiring every field while the endpoint and proxy
+// address stay immutable.
+//
+// Every field is required by the schema, so none is presence-sensitive.
+// AccessLogRetentionDays in particular must stay required: the caller receives
+// a zero-valued Settings, and UpdateSettings copies each field onto the stored
+// row unconditionally, so an omitted value would be written as 0 — which the
+// API documents as "keep indefinitely". Making retention optional would
+// therefore let a client silently maximise log retention by leaving it out.
 func (s *Settings) FromAPIRequest(req *api.AgentNetworkSettingsRequest) {
-	if req.Cluster != nil {
-		s.Cluster = strings.TrimSpace(*req.Cluster)
-	}
+	s.Domain = req.Endpoint
+	s.ProxyAddress = req.ProxyAddress
 	s.EnableLogCollection = req.EnableLogCollection
 	s.EnablePromptCollection = req.EnablePromptCollection
 	s.RedactPii = req.RedactPii
+	s.AccessLogRetentionDays = req.AccessLogRetentionDays
+}
+
+// FromAPICreateRequest applies the optional collection toggles of a bootstrap
+// request onto the receiver (typically DefaultSettings), leaving defaults in
+// place for omitted fields. The identity fields are resolved by the manager
+// from the request's proxy_address / endpoint, not copied here.
+func (s *Settings) FromAPICreateRequest(req *api.AgentNetworkSettingsCreateRequest) {
+	if req.EnableLogCollection != nil {
+		s.EnableLogCollection = *req.EnableLogCollection
+	}
+	if req.EnablePromptCollection != nil {
+		s.EnablePromptCollection = *req.EnablePromptCollection
+	}
+	if req.RedactPii != nil {
+		s.RedactPii = *req.RedactPii
+	}
 	if req.AccessLogRetentionDays != nil {
 		s.AccessLogRetentionDays = *req.AccessLogRetentionDays
 	}
 }
+
+// maxHostnameLength is the RFC 1035 bound on a full domain name.
+const maxHostnameLength = 253
+
+// NormalizeHostname lowercases and trims a caller-supplied hostname and
+// validates its shape: non-empty DNS labels of letters, digits and inner
+// hyphens, joined by single dots, within length bounds. Shapes that
+// canonicalization cannot repair — leading/trailing dots, empty labels,
+// whitespace inside the name — are rejected rather than guessed at, because
+// the value lands in an immutable column.
+func NormalizeHostname(raw string) (string, error) {
+	hostname := strings.ToLower(strings.TrimSpace(raw))
+	if hostname == "" {
+		return "", fmt.Errorf("hostname is empty")
+	}
+	if len(hostname) > maxHostnameLength {
+		return "", fmt.Errorf("hostname exceeds %d characters", maxHostnameLength)
+	}
+	for _, label := range strings.Split(hostname, ".") {
+		if err := validateHostnameLabel(label); err != nil {
+			return "", fmt.Errorf("invalid hostname %q: %w", hostname, err)
+		}
+	}
+	return hostname, nil
+}
+
+func validateHostnameLabel(label string) error {
+	if label == "" {
+		return fmt.Errorf("empty label (leading, trailing or doubled dot)")
+	}
+	if len(label) > 63 {
+		return fmt.Errorf("label %q exceeds 63 characters", label)
+	}
+	if label[0] == '-' || label[len(label)-1] == '-' {
+		return fmt.Errorf("label %q must not start or end with a hyphen", label)
+	}
+	for _, r := range label {
+		switch {
+		case r >= 'a' && r <= 'z':
+		case r >= '0' && r <= '9':
+		case r == '-':
+		default:
+			return fmt.Errorf("label %q contains invalid character %q", label, r)
+		}
+	}
+	return nil
+}

management/internals/modules/agentnetwork/wire_shape_test.go

@@ -5,7 +5,7 @@ import (
 	"encoding/json"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/modules/peers/ephemeral/manager/ephemeral_test.go

@@ -7,7 +7,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
 

management/internals/modules/peers/manager.go

@@ -1,6 +1,6 @@
 package peers
 
-//go:generate go run github.com/golang/mock/mockgen -package peers -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//go:generate go tool mockgen -package peers -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
 
 import (
 	"context"

management/internals/modules/peers/manager_mock.go

@@ -1,5 +1,10 @@
 // Code generated by MockGen. DO NOT EDIT.
 // Source: ./manager.go
+//
+// Generated by this command:
+//
+//	mockgen -package peers -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//
 
 // Package peers is a generated GoMock package.
 package peers
@@ -9,18 +14,19 @@ import (
 	net "net"
 	reflect "reflect"
 
-	gomock "github.com/golang/mock/gomock"
 	network_map "github.com/netbirdio/netbird/management/internals/controllers/network_map"
 	account "github.com/netbirdio/netbird/management/server/account"
 	integrated_validator "github.com/netbirdio/netbird/management/server/integrations/integrated_validator"
 	peer "github.com/netbirdio/netbird/management/server/peer"
 	types "github.com/netbirdio/netbird/management/server/types"
+	gomock "go.uber.org/mock/gomock"
 )
 
 // MockManager is a mock of Manager interface.
 type MockManager struct {
 	ctrl     *gomock.Controller
 	recorder *MockManagerMockRecorder
+	isgomock struct{}
 }
 
 // MockManagerMockRecorder is the mock recorder for MockManager.
@@ -49,7 +55,7 @@ func (m *MockManager) CreateProxyPeer(ctx context.Context, accountID, peerKey, c
 }
 
 // CreateProxyPeer indicates an expected call of CreateProxyPeer.
-func (mr *MockManagerMockRecorder) CreateProxyPeer(ctx, accountID, peerKey, cluster interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) CreateProxyPeer(ctx, accountID, peerKey, cluster any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CreateProxyPeer", reflect.TypeOf((*MockManager)(nil).CreateProxyPeer), ctx, accountID, peerKey, cluster)
 }
@@ -63,7 +69,7 @@ func (m *MockManager) DeletePeers(ctx context.Context, accountID string, peerIDs
 }
 
 // DeletePeers indicates an expected call of DeletePeers.
-func (mr *MockManagerMockRecorder) DeletePeers(ctx, accountID, peerIDs, userID, checkConnected interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) DeletePeers(ctx, accountID, peerIDs, userID, checkConnected any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeletePeers", reflect.TypeOf((*MockManager)(nil).DeletePeers), ctx, accountID, peerIDs, userID, checkConnected)
 }
@@ -78,7 +84,7 @@ func (m *MockManager) GetAllPeers(ctx context.Context, accountID, userID string)
 }
 
 // GetAllPeers indicates an expected call of GetAllPeers.
-func (mr *MockManagerMockRecorder) GetAllPeers(ctx, accountID, userID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetAllPeers(ctx, accountID, userID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAllPeers", reflect.TypeOf((*MockManager)(nil).GetAllPeers), ctx, accountID, userID)
 }
@@ -93,7 +99,7 @@ func (m *MockManager) GetPeer(ctx context.Context, accountID, userID, peerID str
 }
 
 // GetPeer indicates an expected call of GetPeer.
-func (mr *MockManagerMockRecorder) GetPeer(ctx, accountID, userID, peerID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPeer(ctx, accountID, userID, peerID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeer", reflect.TypeOf((*MockManager)(nil).GetPeer), ctx, accountID, userID, peerID)
 }
@@ -108,7 +114,7 @@ func (m *MockManager) GetPeerAccountID(ctx context.Context, peerID string) (stri
 }
 
 // GetPeerAccountID indicates an expected call of GetPeerAccountID.
-func (mr *MockManagerMockRecorder) GetPeerAccountID(ctx, peerID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPeerAccountID(ctx, peerID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeerAccountID", reflect.TypeOf((*MockManager)(nil).GetPeerAccountID), ctx, peerID)
 }
@@ -123,7 +129,7 @@ func (m *MockManager) GetPeerByTunnelIP(ctx context.Context, accountID string, i
 }
 
 // GetPeerByTunnelIP indicates an expected call of GetPeerByTunnelIP.
-func (mr *MockManagerMockRecorder) GetPeerByTunnelIP(ctx, accountID, ip interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPeerByTunnelIP(ctx, accountID, ip any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeerByTunnelIP", reflect.TypeOf((*MockManager)(nil).GetPeerByTunnelIP), ctx, accountID, ip)
 }
@@ -138,7 +144,7 @@ func (m *MockManager) GetPeerID(ctx context.Context, peerKey string) (string, er
 }
 
 // GetPeerID indicates an expected call of GetPeerID.
-func (mr *MockManagerMockRecorder) GetPeerID(ctx, peerKey interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPeerID(ctx, peerKey any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeerID", reflect.TypeOf((*MockManager)(nil).GetPeerID), ctx, peerKey)
 }
@@ -154,7 +160,7 @@ func (m *MockManager) GetPeerWithGroups(ctx context.Context, accountID, peerID s
 }
 
 // GetPeerWithGroups indicates an expected call of GetPeerWithGroups.
-func (mr *MockManagerMockRecorder) GetPeerWithGroups(ctx, accountID, peerID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPeerWithGroups(ctx, accountID, peerID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeerWithGroups", reflect.TypeOf((*MockManager)(nil).GetPeerWithGroups), ctx, accountID, peerID)
 }
@@ -169,7 +175,7 @@ func (m *MockManager) GetPeersByGroupIDs(ctx context.Context, accountID string,
 }
 
 // GetPeersByGroupIDs indicates an expected call of GetPeersByGroupIDs.
-func (mr *MockManagerMockRecorder) GetPeersByGroupIDs(ctx, accountID, groupsIDs interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPeersByGroupIDs(ctx, accountID, groupsIDs any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPeersByGroupIDs", reflect.TypeOf((*MockManager)(nil).GetPeersByGroupIDs), ctx, accountID, groupsIDs)
 }
@@ -181,7 +187,7 @@ func (m *MockManager) SetAccountManager(accountManager account.Manager) {
 }
 
 // SetAccountManager indicates an expected call of SetAccountManager.
-func (mr *MockManagerMockRecorder) SetAccountManager(accountManager interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SetAccountManager(accountManager any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetAccountManager", reflect.TypeOf((*MockManager)(nil).SetAccountManager), accountManager)
 }
@@ -193,7 +199,7 @@ func (m *MockManager) SetIntegratedPeerValidator(integratedPeerValidator integra
 }
 
 // SetIntegratedPeerValidator indicates an expected call of SetIntegratedPeerValidator.
-func (mr *MockManagerMockRecorder) SetIntegratedPeerValidator(integratedPeerValidator interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SetIntegratedPeerValidator(integratedPeerValidator any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetIntegratedPeerValidator", reflect.TypeOf((*MockManager)(nil).SetIntegratedPeerValidator), integratedPeerValidator)
 }
@@ -205,7 +211,7 @@ func (m *MockManager) SetNetworkMapController(networkMapController network_map.C
 }
 
 // SetNetworkMapController indicates an expected call of SetNetworkMapController.
-func (mr *MockManagerMockRecorder) SetNetworkMapController(networkMapController interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SetNetworkMapController(networkMapController any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetNetworkMapController", reflect.TypeOf((*MockManager)(nil).SetNetworkMapController), networkMapController)
 }

management/internals/modules/reverseproxy/accesslogs/manager/manager_test.go

@@ -5,7 +5,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/modules/reverseproxy/activity/interface.go

@@ -0,0 +1,25 @@
+// Package activity records that a principal used a reverse proxy service, so
+// that activity accounting counts people and devices which reach services
+// through the proxy but never touch the dashboard or the management API.
+package activity
+
+import (
+	"context"
+
+	"github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// Manager records reverse proxy usage against the timestamps activity
+// accounting reads. Both methods are best effort from the caller's point of
+// view: a lost record is corrected by the next request, and no authorization
+// decision reads them back.
+type Manager interface {
+	// RecordUserLogin records a completed SSO sign-in to a proxied service.
+	// Service users have no interactive login and are ignored.
+	RecordUserLogin(ctx context.Context, accountID string, user *types.User) error
+	// RecordPeerSeen records that a peer reached a private service over the
+	// mesh, which is what lets its owner count as active. Peers activity
+	// accounting excludes, and peers already seen recently, are ignored.
+	RecordPeerSeen(ctx context.Context, accountID string, peer *peer.Peer) error
+}

management/internals/modules/reverseproxy/activity/manager/manager.go

@@ -0,0 +1,66 @@
+package manager
+
+import (
+	"context"
+	"time"
+
+	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/activity"
+	"github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// peerSeenInterval is how stale a peer's LastSeen must be before reaching a
+// private service refreshes it. Positive tunnel validations are cached on the
+// proxy for five minutes, so without a floor a busy peer would rewrite its row
+// behind every request; an hour still sits well inside the window activity
+// accounting asks about.
+const peerSeenInterval = time.Hour
+
+type managerImpl struct {
+	store store.Store
+}
+
+// NewManager returns the activity manager backed by the management store.
+func NewManager(store store.Store) activity.Manager {
+	return &managerImpl{store: store}
+}
+
+// RecordUserLogin stamps the login the same way the dashboard and device login
+// paths do, so a person who only ever reaches proxied services still has a
+// login on record.
+func (m *managerImpl) RecordUserLogin(ctx context.Context, accountID string, user *types.User) error {
+	if user == nil || user.IsServiceUser {
+		return nil
+	}
+
+	return m.store.SaveUserLastLogin(ctx, accountID, user.Id, time.Now().UTC())
+}
+
+// RecordPeerSeen stamps LastSeen, the column a peer activates its owner
+// through. The peer the caller already holds answers the throttle without a
+// query, so a peer seen inside the interval costs nothing to skip; the same
+// cutoff goes to the store, which enforces it inside the UPDATE so concurrent
+// requests for one peer cannot each write off their own stale read.
+func (m *managerImpl) RecordPeerSeen(ctx context.Context, accountID string, peer *peer.Peer) error {
+	if peer == nil || !countsTowardActivity(peer) {
+		return nil
+	}
+
+	staleBefore := time.Now().UTC().Add(-peerSeenInterval)
+	if peer.Status != nil && peer.Status.LastSeen.After(staleBefore) {
+		return nil
+	}
+
+	_, err := m.store.RefreshPeerLastSeen(ctx, accountID, peer.ID, staleBefore)
+
+	return err
+}
+
+// countsTowardActivity reports whether the peer represents a device a person
+// actually runs. Embedded proxy peers are infrastructure and browser (WASM)
+// clients are ephemeral sessions, so activity accounting ignores both and a
+// write for them could never count.
+func countsTowardActivity(peer *peer.Peer) bool {
+	return !peer.ProxyMeta.Embedded && peer.Meta.KernelVersion != "wasm"
+}

management/internals/modules/reverseproxy/activity/manager/manager_test.go

@@ -0,0 +1,149 @@
+package manager
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// recordingStore captures the two writes the activity manager makes. The
+// embedded interface satisfies the rest and panics if anything else is called,
+// which keeps the manager honest about its surface.
+type recordingStore struct {
+	store.Store
+	logins []loginWrite
+	seen   []seenWrite
+}
+
+type loginWrite struct {
+	accountID string
+	userID    string
+	at        time.Time
+}
+
+type seenWrite struct {
+	accountID   string
+	peerID      string
+	staleBefore time.Time
+}
+
+func (s *recordingStore) SaveUserLastLogin(_ context.Context, accountID, userID string, lastLogin time.Time) error {
+	s.logins = append(s.logins, loginWrite{accountID: accountID, userID: userID, at: lastLogin})
+	return nil
+}
+
+func (s *recordingStore) RefreshPeerLastSeen(_ context.Context, accountID, peerID string, staleBefore time.Time) (bool, error) {
+	s.seen = append(s.seen, seenWrite{accountID: accountID, peerID: peerID, staleBefore: staleBefore})
+	return true, nil
+}
+
+func TestRecordUserLogin(t *testing.T) {
+	tests := []struct {
+		name        string
+		user        *types.User
+		expectWrite bool
+	}{
+		{
+			name:        "regular user is recorded",
+			user:        &types.User{Id: "user1", AccountID: "account1"},
+			expectWrite: true,
+		},
+		{
+			// Activity accounting never counts service users, so a row for one
+			// would be noise.
+			name:        "service user is ignored",
+			user:        &types.User{Id: "svc1", AccountID: "account1", IsServiceUser: true},
+			expectWrite: false,
+		},
+		{
+			name:        "missing user is ignored",
+			user:        nil,
+			expectWrite: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			st := &recordingStore{}
+			require.NoError(t, NewManager(st).RecordUserLogin(context.Background(), "account1", tt.user))
+
+			if !tt.expectWrite {
+				assert.Empty(t, st.logins, "no login should have been recorded")
+				return
+			}
+
+			require.Len(t, st.logins, 1, "exactly one login should have been recorded")
+			assert.Equal(t, "account1", st.logins[0].accountID, "login must be recorded against the service account")
+			assert.Equal(t, tt.user.Id, st.logins[0].userID, "login must be recorded against the signing-in user")
+			assert.Equal(t, time.UTC, st.logins[0].at.Location(), "timestamps are written in UTC")
+			assert.WithinDuration(t, time.Now().UTC(), st.logins[0].at, time.Minute, "login should be stamped now")
+		})
+	}
+}
+
+func TestRecordPeerSeen(t *testing.T) {
+	tests := []struct {
+		name        string
+		peer        *peer.Peer
+		expectWrite bool
+	}{
+		{
+			name:        "peer seen long ago is recorded",
+			peer:        &peer.Peer{ID: "peer1", Status: &peer.PeerStatus{LastSeen: time.Now().Add(-3 * time.Hour)}},
+			expectWrite: true,
+		},
+		{
+			name:        "peer never seen is recorded",
+			peer:        &peer.Peer{ID: "peer1", Status: &peer.PeerStatus{}},
+			expectWrite: true,
+		},
+		{
+			// The throttle. The caller already holds the peer, so skipping a
+			// recently seen one costs nothing.
+			name:        "peer seen inside the interval is skipped",
+			peer:        &peer.Peer{ID: "peer1", Status: &peer.PeerStatus{LastSeen: time.Now().Add(-10 * time.Minute)}},
+			expectWrite: false,
+		},
+		{
+			name:        "embedded proxy peer is skipped",
+			peer:        &peer.Peer{ID: "peer1", ProxyMeta: peer.ProxyMeta{Embedded: true}, Status: &peer.PeerStatus{LastSeen: time.Now().Add(-3 * time.Hour)}},
+			expectWrite: false,
+		},
+		{
+			name:        "browser client is skipped",
+			peer:        &peer.Peer{ID: "peer1", Meta: peer.PeerSystemMeta{KernelVersion: "wasm"}, Status: &peer.PeerStatus{LastSeen: time.Now().Add(-3 * time.Hour)}},
+			expectWrite: false,
+		},
+		{
+			name:        "missing peer is ignored",
+			peer:        nil,
+			expectWrite: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			st := &recordingStore{}
+			require.NoError(t, NewManager(st).RecordPeerSeen(context.Background(), "account1", tt.peer))
+
+			if !tt.expectWrite {
+				assert.Empty(t, st.seen, "no activity should have been recorded")
+				return
+			}
+
+			require.Len(t, st.seen, 1, "exactly one activity write should have been recorded")
+			assert.Equal(t, "account1", st.seen[0].accountID, "activity must be recorded against the service account")
+			assert.Equal(t, tt.peer.ID, st.seen[0].peerID, "activity must be recorded against the calling peer")
+			assert.Equal(t, time.UTC, st.seen[0].staleBefore.Location(), "cutoffs are passed in UTC")
+			assert.WithinDuration(t, time.Now().UTC().Add(-peerSeenInterval), st.seen[0].staleBefore, time.Minute,
+				"the store must enforce the same interval the local check applies")
+		})
+	}
+}

management/internals/modules/reverseproxy/domain/manager/manager.go

@@ -2,24 +2,28 @@ package manager
 
 import (
 	"context"
+	"errors"
 	"fmt"
 	"net"
 	"strings"
 
 	log "github.com/sirupsen/logrus"
 
+	agentnetworkTypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/domain"
 	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/activity"
 	"github.com/netbirdio/netbird/management/server/permissions"
 	"github.com/netbirdio/netbird/management/server/permissions/modules"
 	"github.com/netbirdio/netbird/management/server/permissions/operations"
+	nbstore "github.com/netbirdio/netbird/management/server/store"
 	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/shared/management/status"
 )
 
 type store interface {
 	GetAccount(ctx context.Context, accountID string) (*types.Account, error)
+	GetAgentNetworkSettings(ctx context.Context, lockStrength nbstore.LockingStrength, accountID string) (*agentnetworkTypes.Settings, error)
 
 	GetCustomDomain(ctx context.Context, accountID string, domainID string) (*domain.Domain, error)
 	ListFreeDomains(ctx context.Context, accountID string) ([]string, error)
@@ -311,17 +315,21 @@ func (m Manager) getClusterAllowList(ctx context.Context, accountID string) ([]s
 	if err != nil {
 		return nil, fmt.Errorf("get public cluster addresses: %w", err)
 	}
+	reserved, err := m.reservedGatewayAddress(ctx, accountID)
+	if err != nil {
+		return nil, err
+	}
 	seen := make(map[string]struct{}, len(byopAddresses)+len(publicAddresses))
 	merged := make([]string, 0, len(byopAddresses)+len(publicAddresses))
 	for _, addr := range byopAddresses {
-		if _, ok := seen[addr]; ok {
+		if _, ok := seen[addr]; ok || addr == reserved {
 			continue
 		}
 		seen[addr] = struct{}{}
 		merged = append(merged, addr)
 	}
 	for _, addr := range publicAddresses {
-		if _, ok := seen[addr]; ok {
+		if _, ok := seen[addr]; ok || addr == reserved {
 			continue
 		}
 		seen[addr] = struct{}{}
@@ -330,6 +338,31 @@ func (m Manager) getClusterAllowList(ctx context.Context, accountID string) ([]s
 	return merged, nil
 }
 
+// reservedGatewayAddress returns the account's agent-network gateway address
+// when its settings pin is self-addressed — a proxy dedicated to serving
+// exactly the gateway. Dropping that address from the cluster allow list keeps
+// it from being offered as a cluster for ordinary services, and because the
+// free-domain suffix match is depth-independent, dropping the address rejects
+// every name beneath it as well as the bare one. Only the account's own
+// gateway address can ever appear in its allow list (another tenant's gateway
+// proxy is account-scoped to them), so this single-address exclusion is
+// sufficient. Returns "" when the account has no settings row or a labeled
+// (shared-cluster) pin.
+func (m Manager) reservedGatewayAddress(ctx context.Context, accountID string) (string, error) {
+	settings, err := m.store.GetAgentNetworkSettings(ctx, nbstore.LockingStrengthNone, accountID)
+	if err != nil {
+		var sErr *status.Error
+		if errors.As(err, &sErr) && sErr.Type() == status.NotFound {
+			return "", nil
+		}
+		return "", fmt.Errorf("get agent network settings: %w", err)
+	}
+	if settings == nil || !settings.Dedicated() {
+		return "", nil
+	}
+	return settings.ProxyAddress, nil
+}
+
 func extractClusterFromCustomDomains(serviceDomain string, customDomains []*domain.Domain) (string, bool) {
 	bestCluster := ""
 	bestLen := -1

management/internals/modules/reverseproxy/domain/manager/manager_test.go

@@ -7,6 +7,12 @@ import (
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+
+	agentnetworkTypes "github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/domain"
+	nbstore "github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/types"
+	"github.com/netbirdio/netbird/shared/management/status"
 )
 
 type mockProxyManager struct {
@@ -55,7 +61,7 @@ func TestGetClusterAllowList_BYOPMergedWithPublic(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.NoError(t, err)
 	assert.Equal(t, []string{"byop.example.com", "eu.proxy.netbird.io"}, result)
@@ -71,7 +77,7 @@ func TestGetClusterAllowList_DeduplicatesBYOPAndPublic(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.NoError(t, err)
 	assert.Equal(t, []string{"shared.example.com", "byop.example.com", "eu.proxy.netbird.io"}, result)
@@ -87,7 +93,7 @@ func TestGetClusterAllowList_NoBYOP_FallbackToShared(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.NoError(t, err)
 	assert.Equal(t, []string{"eu.proxy.netbird.io", "us.proxy.netbird.io"}, result)
@@ -100,7 +106,7 @@ func TestGetClusterAllowList_BYOPError_ReturnsError(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.Error(t, err)
 	assert.Nil(t, result)
@@ -117,7 +123,7 @@ func TestGetClusterAllowList_PublicError_ReturnsError(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.Error(t, err)
 	assert.Nil(t, result)
@@ -134,7 +140,7 @@ func TestGetClusterAllowList_BYOPEmptySlice_FallbackToShared(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.NoError(t, err)
 	assert.Equal(t, []string{"eu.proxy.netbird.io"}, result)
@@ -150,8 +156,138 @@ func TestGetClusterAllowList_PublicEmpty_BYOPOnly(t *testing.T) {
 		},
 	}
 
-	mgr := Manager{proxyManager: pm}
+	mgr := Manager{store: &stubStore{}, proxyManager: pm}
 	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
 	require.NoError(t, err)
 	assert.Equal(t, []string{"byop.example.com"}, result)
 }
+
+// stubStore satisfies the manager's narrow store interface for allow-list
+// tests. Only the agent-network settings lookup participates; the default (a
+// nil func) reads as "no settings row", the state most accounts are in.
+type stubStore struct {
+	getAgentNetworkSettingsFunc func(ctx context.Context, accountID string) (*agentnetworkTypes.Settings, error)
+}
+
+func (s *stubStore) GetAccount(context.Context, string) (*types.Account, error) {
+	panic("not used in allow-list tests")
+}
+
+func (s *stubStore) GetAgentNetworkSettings(ctx context.Context, _ nbstore.LockingStrength, accountID string) (*agentnetworkTypes.Settings, error) {
+	if s.getAgentNetworkSettingsFunc != nil {
+		return s.getAgentNetworkSettingsFunc(ctx, accountID)
+	}
+	return nil, status.Errorf(status.NotFound, "agent network settings for account %s not found", accountID)
+}
+
+func (s *stubStore) GetCustomDomain(context.Context, string, string) (*domain.Domain, error) {
+	panic("not used in allow-list tests")
+}
+
+func (s *stubStore) ListFreeDomains(context.Context, string) ([]string, error) {
+	panic("not used in allow-list tests")
+}
+
+func (s *stubStore) ListCustomDomains(context.Context, string) ([]*domain.Domain, error) {
+	panic("not used in allow-list tests")
+}
+
+func (s *stubStore) CreateCustomDomain(context.Context, string, string, string, bool) (*domain.Domain, error) {
+	panic("not used in allow-list tests")
+}
+
+func (s *stubStore) UpdateCustomDomain(context.Context, string, *domain.Domain) (*domain.Domain, error) {
+	panic("not used in allow-list tests")
+}
+
+func (s *stubStore) DeleteCustomDomain(context.Context, string, string) error {
+	panic("not used in allow-list tests")
+}
+
+// TestGetClusterAllowList_DedicatedGatewayAddressExcluded pins invariant (B)'s
+// chokepoint: a self-addressed settings pin reserves the account's gateway
+// address, so it is dropped from the allow list — which, because the
+// free-domain suffix match is depth-independent, rejects every name beneath
+// it as well as the bare one. Other addresses are unaffected.
+func TestGetClusterAllowList_DedicatedGatewayAddressExcluded(t *testing.T) {
+	pm := &mockProxyManager{
+		getActiveClusterAddressesForAccountFunc: func(_ context.Context, _ string) ([]string, error) {
+			return []string{"brave-otter.gateway.example.com", "byop.example.com"}, nil
+		},
+		getActiveClusterAddressesFunc: func(_ context.Context) ([]string, error) {
+			return []string{"eu.proxy.netbird.io"}, nil
+		},
+	}
+	st := &stubStore{
+		getAgentNetworkSettingsFunc: func(_ context.Context, accountID string) (*agentnetworkTypes.Settings, error) {
+			assert.Equal(t, "acc-123", accountID,
+				"the exclusion must look up the requesting account's own settings")
+			return &agentnetworkTypes.Settings{
+				AccountID:    accountID,
+				Domain:       "brave-otter.gateway.example.com",
+				ProxyAddress: "brave-otter.gateway.example.com",
+			}, nil
+		},
+	}
+
+	mgr := Manager{store: st, proxyManager: pm}
+	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
+	require.NoError(t, err)
+	assert.Equal(t, []string{"byop.example.com", "eu.proxy.netbird.io"}, result,
+		"the dedicated gateway address must be reserved from cluster selection")
+}
+
+// TestGetClusterAllowList_LabeledPinDoesNotExclude pins the counterpart: a
+// labeled pin means the gateway rides on a shared cluster serving ordinary
+// services too, so nothing is reserved.
+func TestGetClusterAllowList_LabeledPinDoesNotExclude(t *testing.T) {
+	pm := &mockProxyManager{
+		getActiveClusterAddressesForAccountFunc: func(_ context.Context, _ string) ([]string, error) {
+			return []string{"byop.example.com"}, nil
+		},
+		getActiveClusterAddressesFunc: func(_ context.Context) ([]string, error) {
+			return []string{"eu.proxy.netbird.io"}, nil
+		},
+	}
+	st := &stubStore{
+		getAgentNetworkSettingsFunc: func(_ context.Context, accountID string) (*agentnetworkTypes.Settings, error) {
+			return &agentnetworkTypes.Settings{
+				AccountID:    accountID,
+				Domain:       "violet.eu.proxy.netbird.io",
+				ProxyAddress: "eu.proxy.netbird.io",
+			}, nil
+		},
+	}
+
+	mgr := Manager{store: st, proxyManager: pm}
+	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
+	require.NoError(t, err)
+	assert.Equal(t, []string{"byop.example.com", "eu.proxy.netbird.io"}, result,
+		"a labeled pin reserves nothing")
+}
+
+// TestGetClusterAllowList_SettingsLookupError_ReturnsError pins that a store
+// outage is surfaced rather than silently treated as "nothing reserved" —
+// failing open here would offer a reserved gateway address for ordinary
+// services.
+func TestGetClusterAllowList_SettingsLookupError_ReturnsError(t *testing.T) {
+	pm := &mockProxyManager{
+		getActiveClusterAddressesForAccountFunc: func(_ context.Context, _ string) ([]string, error) {
+			return []string{"byop.example.com"}, nil
+		},
+		getActiveClusterAddressesFunc: func(_ context.Context) ([]string, error) {
+			return []string{"eu.proxy.netbird.io"}, nil
+		},
+	}
+	st := &stubStore{
+		getAgentNetworkSettingsFunc: func(_ context.Context, _ string) (*agentnetworkTypes.Settings, error) {
+			return nil, status.Errorf(status.Internal, "store outage")
+		},
+	}
+
+	mgr := Manager{store: st, proxyManager: pm}
+	result, err := mgr.getClusterAllowList(context.Background(), "acc-123")
+	require.Error(t, err)
+	assert.Nil(t, result)
+	assert.Contains(t, err.Error(), "agent network settings")
+}

management/internals/modules/reverseproxy/proxy/manager.go

@@ -1,6 +1,6 @@
 package proxy
 
-//go:generate go run github.com/golang/mock/mockgen -package proxy -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//go:generate go tool mockgen -package proxy -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
 
 import (
 	"context"

management/internals/modules/reverseproxy/proxy/manager_mock.go

@@ -1,5 +1,10 @@
 // Code generated by MockGen. DO NOT EDIT.
 // Source: ./manager.go
+//
+// Generated by this command:
+//
+//	mockgen -package proxy -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//
 
 // Package proxy is a generated GoMock package.
 package proxy
@@ -9,14 +14,15 @@ import (
 	reflect "reflect"
 	time "time"
 
-	gomock "github.com/golang/mock/gomock"
 	proto "github.com/netbirdio/netbird/shared/management/proto"
+	gomock "go.uber.org/mock/gomock"
 )
 
 // MockManager is a mock of Manager interface.
 type MockManager struct {
 	ctrl     *gomock.Controller
 	recorder *MockManagerMockRecorder
+	isgomock struct{}
 }
 
 // MockManagerMockRecorder is the mock recorder for MockManager.
@@ -45,25 +51,11 @@ func (m *MockManager) CleanupStale(ctx context.Context, inactivityDuration time.
 }
 
 // CleanupStale indicates an expected call of CleanupStale.
-func (mr *MockManagerMockRecorder) CleanupStale(ctx, inactivityDuration interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) CleanupStale(ctx, inactivityDuration any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CleanupStale", reflect.TypeOf((*MockManager)(nil).CleanupStale), ctx, inactivityDuration)
 }
 
-// ClusterSupportsCustomPorts mocks base method.
-func (m *MockManager) ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "ClusterSupportsCustomPorts", ctx, clusterAddr)
-	ret0, _ := ret[0].(*bool)
-	return ret0
-}
-
-// ClusterSupportsCustomPorts indicates an expected call of ClusterSupportsCustomPorts.
-func (mr *MockManagerMockRecorder) ClusterSupportsCustomPorts(ctx, clusterAddr interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCustomPorts", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCustomPorts), ctx, clusterAddr)
-}
-
 // ClusterRequireSubdomain mocks base method.
 func (m *MockManager) ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool {
 	m.ctrl.T.Helper()
@@ -73,7 +65,7 @@ func (m *MockManager) ClusterRequireSubdomain(ctx context.Context, clusterAddr s
 }
 
 // ClusterRequireSubdomain indicates an expected call of ClusterRequireSubdomain.
-func (mr *MockManagerMockRecorder) ClusterRequireSubdomain(ctx, clusterAddr interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ClusterRequireSubdomain(ctx, clusterAddr any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterRequireSubdomain", reflect.TypeOf((*MockManager)(nil).ClusterRequireSubdomain), ctx, clusterAddr)
 }
@@ -87,11 +79,25 @@ func (m *MockManager) ClusterSupportsCrowdSec(ctx context.Context, clusterAddr s
 }
 
 // ClusterSupportsCrowdSec indicates an expected call of ClusterSupportsCrowdSec.
-func (mr *MockManagerMockRecorder) ClusterSupportsCrowdSec(ctx, clusterAddr interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ClusterSupportsCrowdSec(ctx, clusterAddr any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCrowdSec", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCrowdSec), ctx, clusterAddr)
 }
 
+// ClusterSupportsCustomPorts mocks base method.
+func (m *MockManager) ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "ClusterSupportsCustomPorts", ctx, clusterAddr)
+	ret0, _ := ret[0].(*bool)
+	return ret0
+}
+
+// ClusterSupportsCustomPorts indicates an expected call of ClusterSupportsCustomPorts.
+func (mr *MockManagerMockRecorder) ClusterSupportsCustomPorts(ctx, clusterAddr any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCustomPorts", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCustomPorts), ctx, clusterAddr)
+}
+
 // ClusterSupportsPrivate mocks base method.
 func (m *MockManager) ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool {
 	m.ctrl.T.Helper()
@@ -101,7 +107,7 @@ func (m *MockManager) ClusterSupportsPrivate(ctx context.Context, clusterAddr st
 }
 
 // ClusterSupportsPrivate indicates an expected call of ClusterSupportsPrivate.
-func (mr *MockManagerMockRecorder) ClusterSupportsPrivate(ctx, clusterAddr interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ClusterSupportsPrivate(ctx, clusterAddr any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsPrivate", reflect.TypeOf((*MockManager)(nil).ClusterSupportsPrivate), ctx, clusterAddr)
 }
@@ -116,11 +122,40 @@ func (m *MockManager) Connect(ctx context.Context, proxyID, sessionID, clusterAd
 }
 
 // Connect indicates an expected call of Connect.
-func (mr *MockManagerMockRecorder) Connect(ctx, proxyID, sessionID, clusterAddress, ipAddress, accountID, capabilities interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) Connect(ctx, proxyID, sessionID, clusterAddress, ipAddress, accountID, capabilities any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Connect", reflect.TypeOf((*MockManager)(nil).Connect), ctx, proxyID, sessionID, clusterAddress, ipAddress, accountID, capabilities)
 }
 
+// CountAccountProxies mocks base method.
+func (m *MockManager) CountAccountProxies(ctx context.Context, accountID string) (int64, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "CountAccountProxies", ctx, accountID)
+	ret0, _ := ret[0].(int64)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// CountAccountProxies indicates an expected call of CountAccountProxies.
+func (mr *MockManagerMockRecorder) CountAccountProxies(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CountAccountProxies", reflect.TypeOf((*MockManager)(nil).CountAccountProxies), ctx, accountID)
+}
+
+// DeleteAccountCluster mocks base method.
+func (m *MockManager) DeleteAccountCluster(ctx context.Context, clusterAddress, accountID string) error {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "DeleteAccountCluster", ctx, clusterAddress, accountID)
+	ret0, _ := ret[0].(error)
+	return ret0
+}
+
+// DeleteAccountCluster indicates an expected call of DeleteAccountCluster.
+func (mr *MockManagerMockRecorder) DeleteAccountCluster(ctx, clusterAddress, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteAccountCluster", reflect.TypeOf((*MockManager)(nil).DeleteAccountCluster), ctx, clusterAddress, accountID)
+}
+
 // Disconnect mocks base method.
 func (m *MockManager) Disconnect(ctx context.Context, proxyID, sessionID string) error {
 	m.ctrl.T.Helper()
@@ -130,11 +165,26 @@ func (m *MockManager) Disconnect(ctx context.Context, proxyID, sessionID string)
 }
 
 // Disconnect indicates an expected call of Disconnect.
-func (mr *MockManagerMockRecorder) Disconnect(ctx, proxyID, sessionID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) Disconnect(ctx, proxyID, sessionID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Disconnect", reflect.TypeOf((*MockManager)(nil).Disconnect), ctx, proxyID, sessionID)
 }
 
+// GetAccountProxy mocks base method.
+func (m *MockManager) GetAccountProxy(ctx context.Context, accountID string) (*Proxy, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountProxy", ctx, accountID)
+	ret0, _ := ret[0].(*Proxy)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountProxy indicates an expected call of GetAccountProxy.
+func (mr *MockManagerMockRecorder) GetAccountProxy(ctx, accountID any) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountProxy", reflect.TypeOf((*MockManager)(nil).GetAccountProxy), ctx, accountID)
+}
+
 // GetActiveClusterAddresses mocks base method.
 func (m *MockManager) GetActiveClusterAddresses(ctx context.Context) ([]string, error) {
 	m.ctrl.T.Helper()
@@ -145,11 +195,12 @@ func (m *MockManager) GetActiveClusterAddresses(ctx context.Context) ([]string,
 }
 
 // GetActiveClusterAddresses indicates an expected call of GetActiveClusterAddresses.
-func (mr *MockManagerMockRecorder) GetActiveClusterAddresses(ctx interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetActiveClusterAddresses(ctx any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetActiveClusterAddresses", reflect.TypeOf((*MockManager)(nil).GetActiveClusterAddresses), ctx)
 }
 
+// GetActiveClusterAddressesForAccount mocks base method.
 func (m *MockManager) GetActiveClusterAddressesForAccount(ctx context.Context, accountID string) ([]string, error) {
 	m.ctrl.T.Helper()
 	ret := m.ctrl.Call(m, "GetActiveClusterAddressesForAccount", ctx, accountID)
@@ -158,7 +209,8 @@ func (m *MockManager) GetActiveClusterAddressesForAccount(ctx context.Context, a
 	return ret0, ret1
 }
 
-func (mr *MockManagerMockRecorder) GetActiveClusterAddressesForAccount(ctx, accountID interface{}) *gomock.Call {
+// GetActiveClusterAddressesForAccount indicates an expected call of GetActiveClusterAddressesForAccount.
+func (mr *MockManagerMockRecorder) GetActiveClusterAddressesForAccount(ctx, accountID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetActiveClusterAddressesForAccount", reflect.TypeOf((*MockManager)(nil).GetActiveClusterAddressesForAccount), ctx, accountID)
 }
@@ -172,41 +224,11 @@ func (m *MockManager) Heartbeat(ctx context.Context, p *Proxy) error {
 }
 
 // Heartbeat indicates an expected call of Heartbeat.
-func (mr *MockManagerMockRecorder) Heartbeat(ctx, p interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) Heartbeat(ctx, p any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Heartbeat", reflect.TypeOf((*MockManager)(nil).Heartbeat), ctx, p)
 }
 
-// GetAccountProxy mocks base method.
-func (m *MockManager) GetAccountProxy(ctx context.Context, accountID string) (*Proxy, error) {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "GetAccountProxy", ctx, accountID)
-	ret0, _ := ret[0].(*Proxy)
-	ret1, _ := ret[1].(error)
-	return ret0, ret1
-}
-
-// GetAccountProxy indicates an expected call of GetAccountProxy.
-func (mr *MockManagerMockRecorder) GetAccountProxy(ctx, accountID interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountProxy", reflect.TypeOf((*MockManager)(nil).GetAccountProxy), ctx, accountID)
-}
-
-// CountAccountProxies mocks base method.
-func (m *MockManager) CountAccountProxies(ctx context.Context, accountID string) (int64, error) {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "CountAccountProxies", ctx, accountID)
-	ret0, _ := ret[0].(int64)
-	ret1, _ := ret[1].(error)
-	return ret0, ret1
-}
-
-// CountAccountProxies indicates an expected call of CountAccountProxies.
-func (mr *MockManagerMockRecorder) CountAccountProxies(ctx, accountID interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CountAccountProxies", reflect.TypeOf((*MockManager)(nil).CountAccountProxies), ctx, accountID)
-}
-
 // IsClusterAddressAvailable mocks base method.
 func (m *MockManager) IsClusterAddressAvailable(ctx context.Context, clusterAddress, accountID string) (bool, error) {
 	m.ctrl.T.Helper()
@@ -217,29 +239,16 @@ func (m *MockManager) IsClusterAddressAvailable(ctx context.Context, clusterAddr
 }
 
 // IsClusterAddressAvailable indicates an expected call of IsClusterAddressAvailable.
-func (mr *MockManagerMockRecorder) IsClusterAddressAvailable(ctx, clusterAddress, accountID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) IsClusterAddressAvailable(ctx, clusterAddress, accountID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "IsClusterAddressAvailable", reflect.TypeOf((*MockManager)(nil).IsClusterAddressAvailable), ctx, clusterAddress, accountID)
 }
 
-// DeleteAccountCluster mocks base method.
-func (m *MockManager) DeleteAccountCluster(ctx context.Context, clusterAddress, accountID string) error {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "DeleteAccountCluster", ctx, clusterAddress, accountID)
-	ret0, _ := ret[0].(error)
-	return ret0
-}
-
-// DeleteAccountCluster indicates an expected call of DeleteAccountCluster.
-func (mr *MockManagerMockRecorder) DeleteAccountCluster(ctx, clusterAddress, accountID interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteAccountCluster", reflect.TypeOf((*MockManager)(nil).DeleteAccountCluster), ctx, clusterAddress, accountID)
-}
-
 // MockController is a mock of Controller interface.
 type MockController struct {
 	ctrl     *gomock.Controller
 	recorder *MockControllerMockRecorder
+	isgomock struct{}
 }
 
 // MockControllerMockRecorder is the mock recorder for MockController.
@@ -282,7 +291,7 @@ func (m *MockController) GetProxiesForCluster(clusterAddr string) []string {
 }
 
 // GetProxiesForCluster indicates an expected call of GetProxiesForCluster.
-func (mr *MockControllerMockRecorder) GetProxiesForCluster(clusterAddr interface{}) *gomock.Call {
+func (mr *MockControllerMockRecorder) GetProxiesForCluster(clusterAddr any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetProxiesForCluster", reflect.TypeOf((*MockController)(nil).GetProxiesForCluster), clusterAddr)
 }
@@ -296,7 +305,7 @@ func (m *MockController) RegisterProxyToCluster(ctx context.Context, clusterAddr
 }
 
 // RegisterProxyToCluster indicates an expected call of RegisterProxyToCluster.
-func (mr *MockControllerMockRecorder) RegisterProxyToCluster(ctx, clusterAddr, proxyID interface{}) *gomock.Call {
+func (mr *MockControllerMockRecorder) RegisterProxyToCluster(ctx, clusterAddr, proxyID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "RegisterProxyToCluster", reflect.TypeOf((*MockController)(nil).RegisterProxyToCluster), ctx, clusterAddr, proxyID)
 }
@@ -308,7 +317,7 @@ func (m *MockController) SendServiceUpdateToCluster(ctx context.Context, account
 }
 
 // SendServiceUpdateToCluster indicates an expected call of SendServiceUpdateToCluster.
-func (mr *MockControllerMockRecorder) SendServiceUpdateToCluster(ctx, accountID, update, clusterAddr interface{}) *gomock.Call {
+func (mr *MockControllerMockRecorder) SendServiceUpdateToCluster(ctx, accountID, update, clusterAddr any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SendServiceUpdateToCluster", reflect.TypeOf((*MockController)(nil).SendServiceUpdateToCluster), ctx, accountID, update, clusterAddr)
 }
@@ -322,7 +331,7 @@ func (m *MockController) UnregisterProxyFromCluster(ctx context.Context, cluster
 }
 
 // UnregisterProxyFromCluster indicates an expected call of UnregisterProxyFromCluster.
-func (mr *MockControllerMockRecorder) UnregisterProxyFromCluster(ctx, clusterAddr, proxyID interface{}) *gomock.Call {
+func (mr *MockControllerMockRecorder) UnregisterProxyFromCluster(ctx, clusterAddr, proxyID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "UnregisterProxyFromCluster", reflect.TypeOf((*MockController)(nil).UnregisterProxyFromCluster), ctx, clusterAddr, proxyID)
 }

management/internals/modules/reverseproxy/proxytoken/handler_test.go

@@ -9,7 +9,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"

management/internals/modules/reverseproxy/service/interface.go

@@ -1,6 +1,6 @@
 package service
 
-//go:generate go run github.com/golang/mock/mockgen -package service -destination=interface_mock.go -source=./interface.go -build_flags=-mod=mod
+//go:generate go tool mockgen -package service -destination=interface_mock.go -source=./interface.go -build_flags=-mod=mod
 
 import (
 	"context"

management/internals/modules/reverseproxy/service/interface_mock.go

@@ -1,5 +1,10 @@
 // Code generated by MockGen. DO NOT EDIT.
 // Source: ./interface.go
+//
+// Generated by this command:
+//
+//	mockgen -package service -destination=interface_mock.go -source=./interface.go -build_flags=-mod=mod
+//
 
 // Package service is a generated GoMock package.
 package service
@@ -8,14 +13,15 @@ import (
 	context "context"
 	reflect "reflect"
 
-	gomock "github.com/golang/mock/gomock"
 	proxy "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
+	gomock "go.uber.org/mock/gomock"
 )
 
 // MockManager is a mock of Manager interface.
 type MockManager struct {
 	ctrl     *gomock.Controller
 	recorder *MockManagerMockRecorder
+	isgomock struct{}
 }
 
 // MockManagerMockRecorder is the mock recorder for MockManager.
@@ -45,7 +51,7 @@ func (m *MockManager) CreateService(ctx context.Context, accountID, userID strin
 }
 
 // CreateService indicates an expected call of CreateService.
-func (mr *MockManagerMockRecorder) CreateService(ctx, accountID, userID, service interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) CreateService(ctx, accountID, userID, service any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CreateService", reflect.TypeOf((*MockManager)(nil).CreateService), ctx, accountID, userID, service)
 }
@@ -60,7 +66,7 @@ func (m *MockManager) CreateServiceFromPeer(ctx context.Context, accountID, peer
 }
 
 // CreateServiceFromPeer indicates an expected call of CreateServiceFromPeer.
-func (mr *MockManagerMockRecorder) CreateServiceFromPeer(ctx, accountID, peerID, req interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) CreateServiceFromPeer(ctx, accountID, peerID, req any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CreateServiceFromPeer", reflect.TypeOf((*MockManager)(nil).CreateServiceFromPeer), ctx, accountID, peerID, req)
 }
@@ -74,7 +80,7 @@ func (m *MockManager) DeleteAccountCluster(ctx context.Context, accountID, userI
 }
 
 // DeleteAccountCluster indicates an expected call of DeleteAccountCluster.
-func (mr *MockManagerMockRecorder) DeleteAccountCluster(ctx, accountID, userID, clusterAddress interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) DeleteAccountCluster(ctx, accountID, userID, clusterAddress any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteAccountCluster", reflect.TypeOf((*MockManager)(nil).DeleteAccountCluster), ctx, accountID, userID, clusterAddress)
 }
@@ -88,7 +94,7 @@ func (m *MockManager) DeleteAllServices(ctx context.Context, accountID, userID s
 }
 
 // DeleteAllServices indicates an expected call of DeleteAllServices.
-func (mr *MockManagerMockRecorder) DeleteAllServices(ctx, accountID, userID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) DeleteAllServices(ctx, accountID, userID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteAllServices", reflect.TypeOf((*MockManager)(nil).DeleteAllServices), ctx, accountID, userID)
 }
@@ -102,7 +108,7 @@ func (m *MockManager) DeleteService(ctx context.Context, accountID, userID, serv
 }
 
 // DeleteService indicates an expected call of DeleteService.
-func (mr *MockManagerMockRecorder) DeleteService(ctx, accountID, userID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) DeleteService(ctx, accountID, userID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteService", reflect.TypeOf((*MockManager)(nil).DeleteService), ctx, accountID, userID, serviceID)
 }
@@ -117,7 +123,7 @@ func (m *MockManager) GetAccountServices(ctx context.Context, accountID string)
 }
 
 // GetAccountServices indicates an expected call of GetAccountServices.
-func (mr *MockManagerMockRecorder) GetAccountServices(ctx, accountID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetAccountServices(ctx, accountID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountServices", reflect.TypeOf((*MockManager)(nil).GetAccountServices), ctx, accountID)
 }
@@ -132,7 +138,7 @@ func (m *MockManager) GetAllServices(ctx context.Context, accountID, userID stri
 }
 
 // GetAllServices indicates an expected call of GetAllServices.
-func (mr *MockManagerMockRecorder) GetAllServices(ctx, accountID, userID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetAllServices(ctx, accountID, userID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAllServices", reflect.TypeOf((*MockManager)(nil).GetAllServices), ctx, accountID, userID)
 }
@@ -147,7 +153,7 @@ func (m *MockManager) GetClusters(ctx context.Context, accountID, userID string)
 }
 
 // GetClusters indicates an expected call of GetClusters.
-func (mr *MockManagerMockRecorder) GetClusters(ctx, accountID, userID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetClusters(ctx, accountID, userID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetClusters", reflect.TypeOf((*MockManager)(nil).GetClusters), ctx, accountID, userID)
 }
@@ -162,7 +168,7 @@ func (m *MockManager) GetGlobalServices(ctx context.Context) ([]*Service, error)
 }
 
 // GetGlobalServices indicates an expected call of GetGlobalServices.
-func (mr *MockManagerMockRecorder) GetGlobalServices(ctx interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetGlobalServices(ctx any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetGlobalServices", reflect.TypeOf((*MockManager)(nil).GetGlobalServices), ctx)
 }
@@ -177,7 +183,7 @@ func (m *MockManager) GetService(ctx context.Context, accountID, userID, service
 }
 
 // GetService indicates an expected call of GetService.
-func (mr *MockManagerMockRecorder) GetService(ctx, accountID, userID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetService(ctx, accountID, userID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetService", reflect.TypeOf((*MockManager)(nil).GetService), ctx, accountID, userID, serviceID)
 }
@@ -192,7 +198,7 @@ func (m *MockManager) GetServiceByDomain(ctx context.Context, domain string) (*S
 }
 
 // GetServiceByDomain indicates an expected call of GetServiceByDomain.
-func (mr *MockManagerMockRecorder) GetServiceByDomain(ctx, domain interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetServiceByDomain(ctx, domain any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetServiceByDomain", reflect.TypeOf((*MockManager)(nil).GetServiceByDomain), ctx, domain)
 }
@@ -207,7 +213,7 @@ func (m *MockManager) GetServiceByID(ctx context.Context, accountID, serviceID s
 }
 
 // GetServiceByID indicates an expected call of GetServiceByID.
-func (mr *MockManagerMockRecorder) GetServiceByID(ctx, accountID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetServiceByID(ctx, accountID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetServiceByID", reflect.TypeOf((*MockManager)(nil).GetServiceByID), ctx, accountID, serviceID)
 }
@@ -222,7 +228,7 @@ func (m *MockManager) GetServiceIDByTargetID(ctx context.Context, accountID, res
 }
 
 // GetServiceIDByTargetID indicates an expected call of GetServiceIDByTargetID.
-func (mr *MockManagerMockRecorder) GetServiceIDByTargetID(ctx, accountID, resourceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetServiceIDByTargetID(ctx, accountID, resourceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetServiceIDByTargetID", reflect.TypeOf((*MockManager)(nil).GetServiceIDByTargetID), ctx, accountID, resourceID)
 }
@@ -236,7 +242,7 @@ func (m *MockManager) ReloadAllServicesForAccount(ctx context.Context, accountID
 }
 
 // ReloadAllServicesForAccount indicates an expected call of ReloadAllServicesForAccount.
-func (mr *MockManagerMockRecorder) ReloadAllServicesForAccount(ctx, accountID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ReloadAllServicesForAccount(ctx, accountID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ReloadAllServicesForAccount", reflect.TypeOf((*MockManager)(nil).ReloadAllServicesForAccount), ctx, accountID)
 }
@@ -250,7 +256,7 @@ func (m *MockManager) ReloadService(ctx context.Context, accountID, serviceID st
 }
 
 // ReloadService indicates an expected call of ReloadService.
-func (mr *MockManagerMockRecorder) ReloadService(ctx, accountID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ReloadService(ctx, accountID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ReloadService", reflect.TypeOf((*MockManager)(nil).ReloadService), ctx, accountID, serviceID)
 }
@@ -264,7 +270,7 @@ func (m *MockManager) RenewServiceFromPeer(ctx context.Context, accountID, peerI
 }
 
 // RenewServiceFromPeer indicates an expected call of RenewServiceFromPeer.
-func (mr *MockManagerMockRecorder) RenewServiceFromPeer(ctx, accountID, peerID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) RenewServiceFromPeer(ctx, accountID, peerID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "RenewServiceFromPeer", reflect.TypeOf((*MockManager)(nil).RenewServiceFromPeer), ctx, accountID, peerID, serviceID)
 }
@@ -278,7 +284,7 @@ func (m *MockManager) SetCertificateIssuedAt(ctx context.Context, accountID, ser
 }
 
 // SetCertificateIssuedAt indicates an expected call of SetCertificateIssuedAt.
-func (mr *MockManagerMockRecorder) SetCertificateIssuedAt(ctx, accountID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SetCertificateIssuedAt(ctx, accountID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetCertificateIssuedAt", reflect.TypeOf((*MockManager)(nil).SetCertificateIssuedAt), ctx, accountID, serviceID)
 }
@@ -292,7 +298,7 @@ func (m *MockManager) SetStatus(ctx context.Context, accountID, serviceID string
 }
 
 // SetStatus indicates an expected call of SetStatus.
-func (mr *MockManagerMockRecorder) SetStatus(ctx, accountID, serviceID, status interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SetStatus(ctx, accountID, serviceID, status any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetStatus", reflect.TypeOf((*MockManager)(nil).SetStatus), ctx, accountID, serviceID, status)
 }
@@ -304,7 +310,7 @@ func (m *MockManager) StartExposeReaper(ctx context.Context) {
 }
 
 // StartExposeReaper indicates an expected call of StartExposeReaper.
-func (mr *MockManagerMockRecorder) StartExposeReaper(ctx interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) StartExposeReaper(ctx any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "StartExposeReaper", reflect.TypeOf((*MockManager)(nil).StartExposeReaper), ctx)
 }
@@ -318,7 +324,7 @@ func (m *MockManager) StopServiceFromPeer(ctx context.Context, accountID, peerID
 }
 
 // StopServiceFromPeer indicates an expected call of StopServiceFromPeer.
-func (mr *MockManagerMockRecorder) StopServiceFromPeer(ctx, accountID, peerID, serviceID interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) StopServiceFromPeer(ctx, accountID, peerID, serviceID any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "StopServiceFromPeer", reflect.TypeOf((*MockManager)(nil).StopServiceFromPeer), ctx, accountID, peerID, serviceID)
 }
@@ -333,7 +339,7 @@ func (m *MockManager) UpdateService(ctx context.Context, accountID, userID strin
 }
 
 // UpdateService indicates an expected call of UpdateService.
-func (mr *MockManagerMockRecorder) UpdateService(ctx, accountID, userID, service interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) UpdateService(ctx, accountID, userID, service any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "UpdateService", reflect.TypeOf((*MockManager)(nil).UpdateService), ctx, accountID, userID, service)
 }

management/internals/modules/reverseproxy/service/manager/l4_port_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/modules/reverseproxy/service/manager/manager_test.go

@@ -8,7 +8,7 @@ import (
 	"time"
 
 	cachestore "github.com/eko/gocache/lib/v4/store"
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"go.opentelemetry.io/otel/metric/noop"

management/internals/modules/zones/manager/manager_test.go

@@ -5,7 +5,7 @@ import (
 	"fmt"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/modules/zones/records/manager/manager_test.go

@@ -4,7 +4,7 @@ import (
 	"context"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/internals/server/boot.go

@@ -27,6 +27,8 @@ import (
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/accesslogs"
 	accesslogsmanager "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/accesslogs/manager"
+	proxyactivity "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/activity"
+	proxyactivitymanager "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/activity/manager"
 	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
 	"github.com/netbirdio/netbird/management/server/activity"
@@ -231,6 +233,7 @@ func (s *BaseServer) ReverseProxyGRPCServer() *nbgrpc.ProxyServiceServer {
 		proxyService := nbgrpc.NewProxyServiceServer(s.AccessLogsManager(), s.ProxyTokenStore(), s.PKCEVerifierStore(), s.proxyOIDCConfig(), s.PeersManager(), s.UsersManager(), s.IdpManager(), s.ProxyManager(), s.Store())
 		s.AfterInit(func(s *BaseServer) {
 			proxyService.SetServiceManager(s.ServiceManager())
+			proxyService.SetActivityManager(s.ProxyActivityManager())
 			proxyService.SetProxyController(s.ServiceProxyController())
 			proxyService.SetAgentNetworkSynthesizer(newAgentNetworkSynthesizer(s.Store()))
 			proxyService.SetAgentNetworkLimitsService(s.AgentNetworkManager())
@@ -290,6 +293,13 @@ func (s *BaseServer) PKCEVerifierStore() *nbgrpc.PKCEVerifierStore {
 	})
 }
 
+// ProxyActivityManager records reverse proxy usage for activity accounting.
+func (s *BaseServer) ProxyActivityManager() proxyactivity.Manager {
+	return Create(s, func() proxyactivity.Manager {
+		return proxyactivitymanager.NewManager(s.Store())
+	})
+}
+
 func (s *BaseServer) AccessLogsManager() accesslogs.Manager {
 	return Create(s, func() accesslogs.Manager {
 		accessLogManager := accesslogsmanager.NewManager(s.Store(), s.PermissionsManager(), s.GeoLocationManager())

management/internals/server/server_resolve_domains_test.go

@@ -5,7 +5,7 @@ import (
 	"errors"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/require"
 
 	nbconfig "github.com/netbirdio/netbird/management/internals/server/config"

management/internals/shared/grpc/conversion.go

@@ -321,7 +321,7 @@ func buildJWTConfig(config *nbconfig.HttpServerConfig, deviceFlowConfig *nbconfi
 
 	return &proto.JWTConfig{
 		Issuer:       issuer,
-		Audience:     audience,
+		Audience:     audience, //nolint:staticcheck
 		Audiences:    audiences,
 		KeysLocation: keysLocation,
 	}

management/internals/shared/grpc/proxy.go

@@ -29,14 +29,15 @@ import (
 
 	"github.com/netbirdio/netbird/shared/management/domain"
 
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/internals/modules/peers"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/accesslogs"
+	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/activity"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
 	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/sessionkey"
 	"github.com/netbirdio/netbird/management/server/idp"
 	"github.com/netbirdio/netbird/management/server/peer"
-	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/management/server/users"
 	proxyauth "github.com/netbirdio/netbird/proxy/auth"
@@ -61,6 +62,17 @@ type ProxyTokenChecker interface {
 	IsProxyAccessTokenValid(ctx context.Context, tokenID string) (bool, error)
 }
 
+// ProxyConnectAuthorizer authorizes a proxy's claim to the cluster address it
+// declares at connect time. Implementations are supplied by integrations; none
+// is installed by default, so every well-formed claim is authorized — the
+// declared address is otherwise only checked for availability. token is nil
+// when the connection carries no proxy access token. A returned status error
+// is sent to the proxy unchanged; any other error is wrapped as
+// PermissionDenied.
+type ProxyConnectAuthorizer interface {
+	AuthorizeProxyConnect(ctx context.Context, token *types.ProxyAccessToken, proxyID, address string) error
+}
+
 // ProxyServiceServer implements the ProxyService gRPC server
 // AgentNetworkSynthesizer produces in-memory reverse-proxy services from
 // Agent Network provider/policy state for the proxy snapshot path; synthesised
@@ -99,6 +111,9 @@ type ProxyServiceServer struct {
 	// and the post-flight consumption write (RecordLLMUsage). Optional — when
 	// nil both RPCs return Unimplemented.
 	agentNetworkLimits AgentNetworkLimitsService
+	// connectAuthorizer authorizes address claims at proxy connect time.
+	// Optional — when nil every well-formed claim is authorized.
+	connectAuthorizer ProxyConnectAuthorizer
 	// ProxyController for service updates and cluster management
 	proxyController proxy.Controller
 
@@ -114,6 +129,9 @@ type ProxyServiceServer struct {
 	// Manager for IdP-enriched user data (may be nil when no IdP is configured)
 	idpManager idp.Manager
 
+	// Manager that records reverse proxy usage for activity accounting
+	activityManager activity.Manager
+
 	// Store for one-time authentication tokens
 	tokenStore *OneTimeTokenStore
 
@@ -236,6 +254,13 @@ func (s *ProxyServiceServer) SetServiceManager(manager rpservice.Manager) {
 	s.serviceManager = manager
 }
 
+// SetActivityManager wires the manager that records reverse proxy usage.
+func (s *ProxyServiceServer) SetActivityManager(manager activity.Manager) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	s.activityManager = manager
+}
+
 // SetAgentNetworkSynthesizer wires the agent-network service synthesiser.
 // Optional — when nil the snapshot path skips agent-network synthesis. The
 // modules layer injects this after both the proxy server and the agent-network
@@ -262,6 +287,23 @@ func (s *ProxyServiceServer) agentNetworkSynthesizer() AgentNetworkSynthesizer {
 	return s.agentNetworkSynth
 }
 
+// SetProxyConnectAuthorizer wires the connect-time address-claim authorizer.
+// Optional — when nil (the default) every well-formed claim is authorized,
+// which is the behavior without the hook. The modules layer injects this
+// after the proxy server is constructed, like the other setters.
+func (s *ProxyServiceServer) SetProxyConnectAuthorizer(authorizer ProxyConnectAuthorizer) {
+	s.mu.Lock()
+	s.connectAuthorizer = authorizer
+	s.mu.Unlock()
+}
+
+// proxyConnectAuthorizer returns the connect authorizer under read lock.
+func (s *ProxyServiceServer) proxyConnectAuthorizer() ProxyConnectAuthorizer {
+	s.mu.RLock()
+	defer s.mu.RUnlock()
+	return s.connectAuthorizer
+}
+
 // CheckLLMPolicyLimits is the pre-flight policy gate the proxy calls before
 // forwarding an LLM request upstream. Delegates to the agent-network selector,
 // which scores applicable policies by remaining headroom and returns the
@@ -446,8 +488,9 @@ func recvSyncInit(stream proto.ProxyService_SyncMappingsServer) (*proto.SyncMapp
 	return init, nil
 }
 
-// validateProxyConnect validates the proxy ID and address, and checks cluster
-// address availability for account-scoped tokens.
+// validateProxyConnect validates the proxy ID and address, checks cluster
+// address availability for account-scoped tokens, and finally consults the
+// connect authorizer (when installed) on the address claim.
 func (s *ProxyServiceServer) validateProxyConnect(proxyID, address string, ctx context.Context) (proxyConnectParams, error) {
 	if proxyID == "" {
 		return proxyConnectParams{}, status.Errorf(codes.InvalidArgument, "proxy_id is required")
@@ -467,6 +510,19 @@ func (s *ProxyServiceServer) validateProxyConnect(proxyID, address string, ctx c
 		}
 	}
 
+	// The authorizer runs last, outside the account-scoped branch, so it also
+	// sees management-wide and token-less connects. PermissionDenied keeps an
+	// authorization rejection distinguishable from the AlreadyExists address
+	// conflict above in proxy logs.
+	if authorizer := s.proxyConnectAuthorizer(); authorizer != nil {
+		if err := authorizer.AuthorizeProxyConnect(ctx, token, proxyID, address); err != nil {
+			if _, ok := status.FromError(err); ok {
+				return proxyConnectParams{}, err
+			}
+			return proxyConnectParams{}, status.Errorf(codes.PermissionDenied, "proxy connect not authorized: %v", err)
+		}
+	}
+
 	return proxyConnectParams{proxyID: proxyID, address: address}, nil
 }
 
@@ -1255,7 +1311,7 @@ func (s *ProxyServiceServer) authenticateHeader(ctx context.Context, serviceID s
 			lastErr = err
 			continue
 		}
-		return true, "header-user", proxyauth.MethodHeader
+		return true, proxyauth.HeaderUserID, proxyauth.MethodHeader
 	}
 
 	if lastErr != nil {
@@ -1579,9 +1635,62 @@ func (s *ProxyServiceServer) ValidateState(state string) (verifier, redirectURL
 	return verifier, redirectURL, nil
 }
 
+// Denied reasons reported to the proxy when access is refused because of the
+// account status of the user behind the request.
+const (
+	deniedReasonPendingApproval = "pending_approval"
+	deniedReasonUserBlocked     = "user_blocked"
+	deniedReasonUserNotFound    = "user_not_found"
+)
+
+var (
+	// ErrUserPendingApproval reports a user whose account still awaits approval
+	// by an administrator and may therefore not hold a proxy session.
+	ErrUserPendingApproval = errors.New("user pending approval")
+
+	// ErrUserBlocked reports a blocked user, who may not hold a proxy session.
+	ErrUserBlocked = errors.New("user blocked")
+
+	errUserUnresolved = errors.New("user could not be resolved")
+)
+
+// checkUserStatus reports whether the user's account status permits reverse
+// proxy access, returning the denied reason for the proxy access log together
+// with the sentinel error callers match on. A user awaiting approval is stored
+// as both pending and blocked, so the pending state is reported first: it is
+// the one an administrator can act on.
+func checkUserStatus(user *types.User) (string, error) {
+	switch {
+	case user == nil:
+		return deniedReasonUserNotFound, errUserUnresolved
+	case user.PendingApproval:
+		return deniedReasonPendingApproval, ErrUserPendingApproval
+	case user.IsBlocked():
+		return deniedReasonUserBlocked, ErrUserBlocked
+	default:
+		return "", nil
+	}
+}
+
+// userStatusDeniedReason returns the denied reason for callers that report a
+// decision rather than an error, and an empty string when the user may proceed.
+func userStatusDeniedReason(user *types.User) string {
+	reason, _ := checkUserStatus(user)
+	return reason
+}
+
+// sameAccount reports whether a user belongs to a service's account. An empty
+// identifier on either side never matches: two unset accounts must not compare
+// equal into a grant.
+func sameAccount(userAccountID, serviceAccountID string) bool {
+	return userAccountID != "" && serviceAccountID != "" && userAccountID == serviceAccountID
+}
+
 // GenerateSessionToken creates a signed session JWT for the given domain and
 // user. The user's group memberships are embedded in the token so policy-aware
 // middlewares on the proxy can authorise without an extra management round-trip.
+// A user the store cannot resolve, or whose account is pending approval or
+// blocked, gets no token at all, so the browser never receives a session cookie.
 func (s *ProxyServiceServer) GenerateSessionToken(ctx context.Context, domain, userID string, method proxyauth.Method) (string, error) {
 	service, err := s.getServiceByDomain(ctx, domain)
 	if err != nil {
@@ -1592,31 +1701,62 @@ func (s *ProxyServiceServer) GenerateSessionToken(ctx context.Context, domain, u
 		return "", fmt.Errorf("no session key configured for domain: %s", domain)
 	}
 
-	var (
-		email      string
-		groupIDs   []string
-		groupNames []string
-	)
-	if s.usersManager != nil {
-		user, userGroups, uerr := s.usersManager.GetUserWithGroups(ctx, userID)
-		if uerr != nil {
-			log.WithContext(ctx).Debugf("session token mint: lookup user %s: %v", userID, uerr)
-		} else if user != nil {
-			email = user.Email
-			groupIDs, groupNames = pairGroupIDsAndNames(userGroups)
-		}
+	if s.usersManager == nil {
+		return "", errors.New("users manager not configured")
 	}
 
-	return sessionkey.SignToken(
+	user, userGroups, err := s.usersManager.GetUserWithGroups(ctx, userID)
+	if err != nil {
+		return "", fmt.Errorf("get user %s: %w", userID, err)
+	}
+
+	if user == nil {
+		return "", fmt.Errorf("get user %s: %w", userID, errUserUnresolved)
+	}
+
+	// Bind the OIDC identity to the service's account before signing anything
+	// with that service's session key. The proxy validates an installed cookie
+	// locally against the service public key, so a token minted for a user of
+	// another account would be honoured without a management round-trip.
+	if !sameAccount(user.AccountID, service.AccountID) {
+		return "", fmt.Errorf("user %s does not belong to the service account", userID)
+	}
+
+	if _, err := checkUserStatus(user); err != nil {
+		return "", fmt.Errorf("session token for user %s: %w", userID, err)
+	}
+
+	groupIDs, groupNames := pairGroupIDsAndNames(userGroups)
+
+	token, err := sessionkey.SignToken(
 		service.SessionPrivateKey,
 		userID,
-		email,
+		user.Email,
 		domain,
 		method,
 		groupIDs,
 		groupNames,
 		proxyauth.DefaultSessionExpiry,
 	)
+	if err != nil {
+		return "", err
+	}
+
+	s.recordUserLogin(ctx, service.AccountID, user)
+
+	return token, nil
+}
+
+// recordUserLogin hands the sign-in to the activity manager. The RPC must not
+// fail on it, so the error is logged and dropped here rather than returned.
+func (s *ProxyServiceServer) recordUserLogin(ctx context.Context, accountID string, user *types.User) {
+	if s.activityManager == nil {
+		return
+	}
+
+	if err := s.activityManager.RecordUserLogin(ctx, accountID, user); err != nil {
+		log.WithContext(ctx).Debugf("record proxy login for user %s: %v", user.Id, err)
+	}
 }
 
 // ValidateUserGroupAccess checks if a user has access to a service.
@@ -1628,6 +1768,10 @@ func (s *ProxyServiceServer) ValidateUserGroupAccess(ctx context.Context, domain
 		return fmt.Errorf("user not found: %s", userID)
 	}
 
+	if _, err := checkUserStatus(user); err != nil {
+		return fmt.Errorf("user %s denied access to domain %s: %w", userID, domain, err)
+	}
+
 	service, err := s.getAccountServiceByDomain(ctx, user.AccountID, domain)
 	if err != nil {
 		return err
@@ -1682,10 +1826,7 @@ func (s *ProxyServiceServer) ValidateSession(ctx context.Context, req *proto.Val
 	sessionToken := req.GetSessionToken()
 
 	if domain == "" || sessionToken == "" {
-		return &proto.ValidateSessionResponse{
-			Valid:        false,
-			DeniedReason: "missing domain or session_token",
-		}, nil
+		return deniedSessionResponse("missing domain or session_token"), nil
 	}
 
 	service, err := s.getServiceByDomain(ctx, domain)
@@ -1695,83 +1836,49 @@ func (s *ProxyServiceServer) ValidateSession(ctx context.Context, req *proto.Val
 			"error":  err.Error(),
 		}).Debug("ValidateSession: service not found")
 		//nolint:nilerr
-		return &proto.ValidateSessionResponse{
-			Valid:        false,
-			DeniedReason: "service_not_found",
-		}, nil
+		return deniedSessionResponse("service_not_found"), nil
 	}
 
 	if err := enforceAccountScope(ctx, service.AccountID); err != nil {
 		return nil, err
 	}
 
-	pubKeyBytes, err := base64.StdEncoding.DecodeString(service.SessionPublicKey)
-	if err != nil {
-		log.WithFields(log.Fields{
-			"domain": domain,
-			"error":  err.Error(),
-		}).Error("ValidateSession: decode public key")
-		//nolint:nilerr
-		return &proto.ValidateSessionResponse{
-			Valid:        false,
-			DeniedReason: "invalid_service_config",
-		}, nil
-	}
-
-	userID, _, _, _, _, err := proxyauth.ValidateSessionJWT(sessionToken, domain, pubKeyBytes)
-	if err != nil {
-		log.WithFields(log.Fields{
-			"domain": domain,
-			"error":  err.Error(),
-		}).Debug("ValidateSession: invalid session token")
-		//nolint:nilerr
-		return &proto.ValidateSessionResponse{
-			Valid:        false,
-			DeniedReason: "invalid_token",
-		}, nil
+	userID, reason := sessionTokenSubject(domain, service, sessionToken)
+	if reason != "" {
+		return deniedSessionResponse(reason), nil
 	}
 
 	user, userGroups, err := s.usersManager.GetUserWithGroups(ctx, userID)
-	if err != nil {
+	if err != nil || user == nil {
 		log.WithFields(log.Fields{
 			"domain":  domain,
 			"user_id": userID,
-			"error":   err.Error(),
+			"error":   err,
 		}).Debug("ValidateSession: user not found")
 		//nolint:nilerr
-		return &proto.ValidateSessionResponse{
-			Valid:        false,
-			DeniedReason: "user_not_found",
-		}, nil
+		return deniedSessionResponse(deniedReasonUserNotFound), nil
 	}
 
-	if user.AccountID != service.AccountID {
+	// A user from another account gets a bare response: none of their identity
+	// belongs in an answer to a proxy serving a different account.
+	if !sameAccount(user.AccountID, service.AccountID) {
 		log.WithFields(log.Fields{
 			"domain":          domain,
 			"user_id":         userID,
 			"user_account":    user.AccountID,
 			"service_account": service.AccountID,
 		}).Debug("ValidateSession: user account mismatch")
-		//nolint:nilerr
-		return &proto.ValidateSessionResponse{
-			Valid:        false,
-			DeniedReason: "account_mismatch",
-		}, nil
+		return deniedSessionResponse("account_mismatch"), nil
 	}
 
-	if err := s.checkGroupAccess(service, user); err != nil {
-		log.WithFields(log.Fields{
-			"domain":  domain,
-			"user_id": userID,
-			"error":   err.Error(),
-		}).Debug("ValidateSession: access denied")
-		groupIDs, groupNames := pairGroupIDsAndNames(userGroups)
-		//nolint:nilerr
+	groupIDs, groupNames := pairGroupIDsAndNames(userGroups)
+
+	if reason := s.accountUserDeniedReason(domain, service, user); reason != "" {
 		return &proto.ValidateSessionResponse{
 			Valid:          false,
 			UserId:         user.Id,
 			UserEmail:      user.Email,
-			DeniedReason:   "not_in_group",
+			DeniedReason:   reason,
 			PeerGroupIds:   groupIDs,
 			PeerGroupNames: groupNames,
 		}, nil
@@ -1783,7 +1890,6 @@ func (s *ProxyServiceServer) ValidateSession(ctx context.Context, req *proto.Val
 		"email":   user.Email,
 	}).Debug("ValidateSession: access granted")
 
-	groupIDs, groupNames := pairGroupIDsAndNames(userGroups)
 	return &proto.ValidateSessionResponse{
 		Valid:          true,
 		UserId:         user.Id,
@@ -1793,6 +1899,66 @@ func (s *ProxyServiceServer) ValidateSession(ctx context.Context, req *proto.Val
 	}, nil
 }
 
+// deniedSessionResponse builds a denial that carries no identity, for the
+// checks that run before a user of this service's account is resolved.
+func deniedSessionResponse(reason string) *proto.ValidateSessionResponse {
+	return &proto.ValidateSessionResponse{
+		Valid:        false,
+		DeniedReason: reason,
+	}
+}
+
+// sessionTokenSubject verifies the session token against the service's session
+// key and returns the user it was minted for, or the reason it cannot be
+// trusted.
+func sessionTokenSubject(domain string, service *rpservice.Service, sessionToken string) (userID, deniedReason string) {
+	pubKeyBytes, err := base64.StdEncoding.DecodeString(service.SessionPublicKey)
+	if err != nil {
+		log.WithFields(log.Fields{
+			"domain": domain,
+			"error":  err.Error(),
+		}).Error("ValidateSession: decode public key")
+		return "", "invalid_service_config"
+	}
+
+	userID, _, _, _, _, err = proxyauth.ValidateSessionJWT(sessionToken, domain, pubKeyBytes)
+	if err != nil {
+		log.WithFields(log.Fields{
+			"domain": domain,
+			"error":  err.Error(),
+		}).Debug("ValidateSession: invalid session token")
+		return "", "invalid_token"
+	}
+
+	return userID, ""
+}
+
+// accountUserDeniedReason gates a user of the service's own account, returning
+// an empty string when access is granted. Account status comes before group
+// membership: a user awaiting approval or blocked has no access regardless of
+// the groups they were auto-assigned.
+func (s *ProxyServiceServer) accountUserDeniedReason(domain string, service *rpservice.Service, user *types.User) string {
+	if reason := userStatusDeniedReason(user); reason != "" {
+		log.WithFields(log.Fields{
+			"domain":  domain,
+			"user_id": user.Id,
+			"reason":  reason,
+		}).Debug("ValidateSession: user status denies access")
+		return reason
+	}
+
+	if err := s.checkGroupAccess(service, user); err != nil {
+		log.WithFields(log.Fields{
+			"domain":  domain,
+			"user_id": user.Id,
+			"error":   err.Error(),
+		}).Debug("ValidateSession: access denied")
+		return "not_in_group"
+	}
+
+	return ""
+}
+
 func (s *ProxyServiceServer) getServiceByDomain(ctx context.Context, domain string) (*rpservice.Service, error) {
 	service, err := s.serviceManager.GetServiceByDomain(ctx, domain)
 	if err == nil {
@@ -1907,7 +2073,20 @@ func (s *ProxyServiceServer) ValidateTunnelPeer(ctx context.Context, req *proto.
 	}
 
 	groupIDs, groupNames := pairGroupIDsAndNames(peerGroups)
-	principalID, displayIdentity := s.getTunnelPeerInfo(ctx, domain, service, peer)
+	owner := s.resolvePeerOwner(ctx, peer, service.AccountID)
+	principalID, displayIdentity := s.getTunnelPeerInfo(ctx, domain, service, peer, owner)
+
+	if reason := peerOwnerDeniedReason(peer, owner); reason != "" {
+		log.WithFields(log.Fields{"domain": domain, "peer_id": peer.ID, "user_id": peer.UserID, "reason": reason}).Debug("ValidateTunnelPeer: owner status denies access")
+		return &proto.ValidateTunnelPeerResponse{
+			Valid:          false,
+			UserId:         principalID,
+			UserEmail:      displayIdentity,
+			DeniedReason:   reason,
+			PeerGroupIds:   groupIDs,
+			PeerGroupNames: groupNames,
+		}, nil
+	}
 
 	if err := checkPeerGroupAccess(service, groupIDs); err != nil {
 		log.WithFields(log.Fields{"domain": domain, "peer_id": peer.ID, "error": err.Error()}).Debug("ValidateTunnelPeer: access denied")
@@ -1927,6 +2106,8 @@ func (s *ProxyServiceServer) ValidateTunnelPeer(ctx context.Context, req *proto.
 		return nil, err
 	}
 
+	s.recordPeerSeen(ctx, service.AccountID, peer)
+
 	log.WithFields(log.Fields{
 		"domain":       domain,
 		"tunnel_ip":    tunnelIPStr,
@@ -1944,9 +2125,67 @@ func (s *ProxyServiceServer) ValidateTunnelPeer(ctx context.Context, req *proto.
 	}, nil
 }
 
+// recordPeerSeen hands the mesh request to the activity manager. The RPC must
+// not fail on it, so the error is logged and dropped here rather than returned.
+func (s *ProxyServiceServer) recordPeerSeen(ctx context.Context, accountID string, peer *peer.Peer) {
+	if s.activityManager == nil {
+		return
+	}
+
+	if err := s.activityManager.RecordPeerSeen(ctx, accountID, peer); err != nil {
+		log.WithContext(ctx).Debugf("record proxy activity for peer %s: %v", peer.ID, err)
+	}
+}
+
+// resolvePeerOwner returns the user a peer is linked to, once per request so
+// the status gate and the identity resolution below share a single lookup.
+// Unlinked peers (machine agents) have no owner. A lookup that fails returns
+// nil rather than an error: both callers treat an unresolved owner the same
+// way, and neither may trust one it could not read.
+func (s *ProxyServiceServer) resolvePeerOwner(ctx context.Context, peer *peer.Peer, accountID string) *types.User {
+	if peer.UserID == "" {
+		return nil
+	}
+
+	user, err := s.usersManager.GetUser(ctx, peer.UserID)
+	if err != nil {
+		log.WithContext(ctx).Debugf("ValidateTunnelPeer: look up owner %s of peer %s: %v", peer.UserID, peer.ID, err)
+		return nil
+	}
+
+	// The lookup is by user ID alone, so a peer row pointing outside the
+	// service's account would otherwise resolve a foreign user. Leave the owner
+	// unresolved instead: the gate denies it, and neither the response nor the
+	// minted token carries an identity from another account.
+	if !sameAccount(user.AccountID, accountID) {
+		log.WithContext(ctx).Debugf("ValidateTunnelPeer: owner %s of peer %s belongs to another account", peer.UserID, peer.ID)
+		return nil
+	}
+
+	return user
+}
+
+// peerOwnerDeniedReason gates the mesh fast-path on the account status of the
+// peer's owning user, so a user blocked after registering a peer loses
+// mesh-origin access too. Unlinked peers (machine agents) have no owner to gate
+// on and stay first-class callers. An owner the store cannot resolve denies:
+// an unavailable lookup must not grant access.
+func peerOwnerDeniedReason(peer *peer.Peer, owner *types.User) string {
+	if peer.UserID == "" {
+		return ""
+	}
+
+	if owner == nil {
+		return deniedReasonUserNotFound
+	}
+
+	return userStatusDeniedReason(owner)
+}
+
 // getTunnelPeerInfo returns the principal ID and display name for a peer, e.g. a
-// user or peer ID, and peer name or user email.
-func (s *ProxyServiceServer) getTunnelPeerInfo(ctx context.Context, domain string, service *rpservice.Service, peer *peer.Peer) (string, string) {
+// user or peer ID, and peer name or user email. owner is the already-resolved
+// user the peer is linked to, or nil.
+func (s *ProxyServiceServer) getTunnelPeerInfo(ctx context.Context, domain string, service *rpservice.Service, peer *peer.Peer, owner *types.User) (string, string) {
 	// Resolve the principal: when the peer is linked to a user, the human is the
 	// principal so multiple peers owned by the same user share a single
 	// identity. Unlinked peers (machine agents) are their own principal keyed on
@@ -1963,10 +2202,10 @@ func (s *ProxyServiceServer) getTunnelPeerInfo(ctx context.Context, domain strin
 	principalID := peer.UserID
 	displayIdentity := peer.Name
 	// Stored column first (cheap, but often empty for OIDC-provisioned users).
-	if user, uerr := s.usersManager.GetUser(ctx, peer.UserID); uerr == nil && user != nil {
-		principalID = user.Id
-		if user.Email != "" {
-			displayIdentity = user.Email
+	if owner != nil {
+		principalID = owner.Id
+		if owner.Email != "" {
+			displayIdentity = owner.Email
 		}
 	}
 	// IdP enrichment wins when available — the stored email column is a

management/internals/shared/grpc/proxy_connect_authorizer_test.go

@@ -0,0 +1,168 @@
+package grpc
+
+import (
+	"context"
+	"errors"
+	"testing"
+
+	"go.uber.org/mock/gomock"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"google.golang.org/grpc/codes"
+	grpcstatus "google.golang.org/grpc/status"
+
+	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
+	"github.com/netbirdio/netbird/management/server/types"
+)
+
+// capturingAuthorizer records the arguments of the last AuthorizeProxyConnect
+// call and returns a fixed error.
+type capturingAuthorizer struct {
+	called  int
+	token   *types.ProxyAccessToken
+	proxyID string
+	address string
+	err     error
+}
+
+func (a *capturingAuthorizer) AuthorizeProxyConnect(_ context.Context, token *types.ProxyAccessToken, proxyID, address string) error {
+	a.called++
+	a.token = token
+	a.proxyID = proxyID
+	a.address = address
+	return a.err
+}
+
+// authorizerServer builds a ProxyServiceServer whose proxy manager reports
+// every cluster address as available, so the authorizer is the only thing
+// standing between a claim and success.
+func authorizerServer(t *testing.T) *ProxyServiceServer {
+	t.Helper()
+	ctrl := gomock.NewController(t)
+	mgr := proxy.NewMockManager(ctrl)
+	mgr.EXPECT().IsClusterAddressAvailable(gomock.Any(), gomock.Any(), gomock.Any()).Return(true, nil).AnyTimes()
+	return &ProxyServiceServer{proxyManager: mgr}
+}
+
+// TestValidateProxyConnect_NilAuthorizerUnchanged guards the no-behavior-change
+// claim: with no authorizer installed — the OSS default — a well-formed claim
+// succeeds and malformed input is rejected exactly as before the hook existed.
+func TestValidateProxyConnect_NilAuthorizerUnchanged(t *testing.T) {
+	s := authorizerServer(t)
+
+	params, err := s.validateProxyConnect("proxy-1", "cluster.example.com", scopedCtx("acc-1"))
+	require.NoError(t, err)
+	assert.Equal(t, "proxy-1", params.proxyID)
+	assert.Equal(t, "cluster.example.com", params.address)
+
+	_, err = s.validateProxyConnect("", "cluster.example.com", scopedCtx("acc-1"))
+	require.Error(t, err)
+	st, ok := grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.InvalidArgument, st.Code(), "missing proxy_id must stay InvalidArgument")
+
+	_, err = s.validateProxyConnect("proxy-1", "not a hostname", scopedCtx("acc-1"))
+	require.Error(t, err)
+	st, ok = grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.InvalidArgument, st.Code(), "invalid address must stay InvalidArgument")
+}
+
+// TestValidateProxyConnect_AuthorizerReceivesClaim pins the hook contract: the
+// authorizer sees the presented token and the claimed proxy ID and address,
+// and an authorized claim proceeds.
+func TestValidateProxyConnect_AuthorizerReceivesClaim(t *testing.T) {
+	s := authorizerServer(t)
+	auth := &capturingAuthorizer{}
+	s.SetProxyConnectAuthorizer(auth)
+
+	params, err := s.validateProxyConnect("proxy-1", "cluster.example.com", scopedCtx("acc-1"))
+	require.NoError(t, err)
+	assert.Equal(t, "cluster.example.com", params.address)
+
+	require.Equal(t, 1, auth.called, "authorizer must be consulted exactly once per connect")
+	assert.Equal(t, "proxy-1", auth.proxyID)
+	assert.Equal(t, "cluster.example.com", auth.address)
+	require.NotNil(t, auth.token, "the presented token must be handed to the authorizer")
+	require.NotNil(t, auth.token.AccountID)
+	assert.Equal(t, "acc-1", *auth.token.AccountID)
+}
+
+// TestValidateProxyConnect_PlainErrorBecomesPermissionDenied pins the error
+// mapping: a non-status error from the authorizer surfaces as
+// PermissionDenied — distinguishable from the AlreadyExists used for address
+// conflicts — and the claim does not proceed even though the address itself
+// was available.
+func TestValidateProxyConnect_PlainErrorBecomesPermissionDenied(t *testing.T) {
+	s := authorizerServer(t)
+	s.SetProxyConnectAuthorizer(&capturingAuthorizer{err: errors.New("not the assigned credential")})
+
+	_, err := s.validateProxyConnect("proxy-1", "cluster.example.com", scopedCtx("acc-1"))
+	require.Error(t, err)
+	st, ok := grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.PermissionDenied, st.Code())
+	assert.Contains(t, st.Message(), "not the assigned credential", "the authorizer's reason must survive into the status message")
+}
+
+// TestValidateProxyConnect_StatusErrorPassesThrough pins that an authorizer
+// which chooses its own status code is not second-guessed.
+func TestValidateProxyConnect_StatusErrorPassesThrough(t *testing.T) {
+	s := authorizerServer(t)
+	s.SetProxyConnectAuthorizer(&capturingAuthorizer{
+		err: grpcstatus.Errorf(codes.ResourceExhausted, "try later"),
+	})
+
+	_, err := s.validateProxyConnect("proxy-1", "cluster.example.com", scopedCtx("acc-1"))
+	require.Error(t, err)
+	st, ok := grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.ResourceExhausted, st.Code(), "a status error must pass through unchanged")
+	assert.Equal(t, "try later", st.Message())
+}
+
+// TestValidateProxyConnect_AuthorizerSeesGlobalAndTokenlessConnects pins the
+// call-site placement: the authorizer sits outside the account-scoped branch,
+// so management-wide tokens (AccountID == nil) and connections without any
+// token are also presented to it rather than bypassing policy.
+func TestValidateProxyConnect_AuthorizerSeesGlobalAndTokenlessConnects(t *testing.T) {
+	s := &ProxyServiceServer{} // no proxy manager: neither path may reach the availability check
+
+	auth := &capturingAuthorizer{}
+	s.SetProxyConnectAuthorizer(auth)
+
+	_, err := s.validateProxyConnect("proxy-1", "cluster.example.com", globalCtx())
+	require.NoError(t, err)
+	require.Equal(t, 1, auth.called, "a management-wide token must still be presented to the authorizer")
+	require.NotNil(t, auth.token)
+	assert.Nil(t, auth.token.AccountID)
+
+	_, err = s.validateProxyConnect("proxy-1", "cluster.example.com", context.Background())
+	require.NoError(t, err)
+	require.Equal(t, 2, auth.called, "a token-less connect must still be presented to the authorizer")
+	assert.Nil(t, auth.token, "no token in context must surface as a nil token, not a zero value")
+}
+
+// TestValidateProxyConnect_AuthorizerRunsLast pins the ordering: input
+// validation and the availability check precede policy, so the authorizer is
+// never consulted about a claim that is malformed or already rejected.
+func TestValidateProxyConnect_AuthorizerRunsLast(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr := proxy.NewMockManager(ctrl)
+	mgr.EXPECT().IsClusterAddressAvailable(gomock.Any(), gomock.Any(), gomock.Any()).Return(false, nil)
+	s := &ProxyServiceServer{proxyManager: mgr}
+
+	auth := &capturingAuthorizer{}
+	s.SetProxyConnectAuthorizer(auth)
+
+	_, err := s.validateProxyConnect("proxy-1", "not a hostname", scopedCtx("acc-1"))
+	require.Error(t, err)
+	assert.Zero(t, auth.called, "a malformed address must be rejected before policy runs")
+
+	_, err = s.validateProxyConnect("proxy-1", "cluster.example.com", scopedCtx("acc-1"))
+	require.Error(t, err)
+	st, ok := grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.AlreadyExists, st.Code(), "an address conflict must keep its own status")
+	assert.Zero(t, auth.called, "a conflicting address must be rejected before policy runs")
+}

management/internals/shared/grpc/proxy_group_access_test.go

@@ -5,6 +5,7 @@ import (
 	"errors"
 	"net"
 	"testing"
+	"time"
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -119,11 +120,13 @@ func (m *mockReverseProxyManager) GetClusters(_ context.Context, _, _ string) ([
 }
 
 type mockUsersManager struct {
-	users map[string]*types.User
-	err   error
+	users        map[string]*types.User
+	err          error
+	getUserCalls int
 }
 
 func (m *mockUsersManager) GetUser(ctx context.Context, userID string) (*types.User, error) {
+	m.getUserCalls++
 	if m.err != nil {
 		return nil, m.err
 	}
@@ -153,6 +156,27 @@ type mockTunnelPeersManager struct {
 	groupsErr error
 }
 
+// mockActivityManager records what the RPC handed to the activity manager. The
+// policy (throttling, exclusions) is the manager's and is tested there; these
+// tests only pin which requests reach it.
+type mockActivityManager struct {
+	seenMarks []seenMark
+}
+
+type seenMark struct {
+	accountID string
+	peerID    string
+}
+
+func (m *mockActivityManager) RecordUserLogin(_ context.Context, _ string, _ *types.User) error {
+	return nil
+}
+
+func (m *mockActivityManager) RecordPeerSeen(_ context.Context, accountID string, peer *peer.Peer) error {
+	m.seenMarks = append(m.seenMarks, seenMark{accountID: accountID, peerID: peer.ID})
+	return nil
+}
+
 func (m *mockTunnelPeersManager) GetPeerByTunnelIP(_ context.Context, _ string, _ net.IP) (*peer.Peer, error) {
 	return m.peer, m.peerErr
 }
@@ -350,6 +374,64 @@ func TestValidateUserGroupAccess(t *testing.T) {
 			},
 			expectErr: false,
 		},
+		{
+			name:   "user pending approval denied despite group membership",
+			domain: "app.example.com",
+			userID: "user1",
+			proxiesByAccount: map[string][]*service.Service{
+				"account1": {{
+					Domain:    "app.example.com",
+					AccountID: "account1",
+					Auth: service.AuthConfig{
+						BearerAuth: &service.BearerAuthConfig{
+							Enabled:            true,
+							DistributionGroups: []string{"group1"},
+						},
+					},
+				}},
+			},
+			users: map[string]*types.User{
+				// The approval flow stores a pending user as blocked as well.
+				"user1": {Id: "user1", AccountID: "account1", AutoGroups: []string{"group1"}, Blocked: true, PendingApproval: true},
+			},
+			expectErr:    true,
+			expectErrMsg: "user pending approval",
+		},
+		{
+			name:   "blocked user denied despite group membership",
+			domain: "app.example.com",
+			userID: "user1",
+			proxiesByAccount: map[string][]*service.Service{
+				"account1": {{
+					Domain:    "app.example.com",
+					AccountID: "account1",
+					Auth: service.AuthConfig{
+						BearerAuth: &service.BearerAuthConfig{
+							Enabled:            true,
+							DistributionGroups: []string{"group1"},
+						},
+					},
+				}},
+			},
+			users: map[string]*types.User{
+				"user1": {Id: "user1", AccountID: "account1", AutoGroups: []string{"group1"}, Blocked: true},
+			},
+			expectErr:    true,
+			expectErrMsg: "user blocked",
+		},
+		{
+			name:   "blocked user denied on a service with no auth configured",
+			domain: "app.example.com",
+			userID: "user1",
+			proxiesByAccount: map[string][]*service.Service{
+				"account1": {{Domain: "app.example.com", AccountID: "account1", Auth: service.AuthConfig{}}},
+			},
+			users: map[string]*types.User{
+				"user1": {Id: "user1", AccountID: "account1", Blocked: true},
+			},
+			expectErr:    true,
+			expectErrMsg: "user blocked",
+		},
 		{
 			name:             "proxy manager error",
 			domain:           "app.example.com",
@@ -421,17 +503,18 @@ func TestValidateTunnelPeerUserEmailEnrichment(t *testing.T) {
 	storedUserNoEmail := map[string]*types.User{userID: {Id: userID, AccountID: accountID, Email: ""}}
 
 	tests := []struct {
-		name         string
-		peerUserID   string
-		storedUsers  map[string]*types.User
-		storedErr    error
-		noIdP        bool
-		idpEmail     string
-		idpHasData   bool
-		idpErr       error
-		expectEmail  string
-		expectUserID string
-		expectIdPHit bool
+		name               string
+		peerUserID         string
+		storedUsers        map[string]*types.User
+		storedErr          error
+		noIdP              bool
+		idpEmail           string
+		idpHasData         bool
+		idpErr             error
+		expectEmail        string
+		expectUserID       string
+		expectIdPHit       bool
+		expectDeniedReason string
 	}{
 		{
 			name:         "idp email wins over stored email",
@@ -490,14 +573,17 @@ func TestValidateTunnelPeerUserEmailEnrichment(t *testing.T) {
 			expectIdPHit: true,
 		},
 		{
-			name:         "idp email when stored user missing keeps peer.UserID as principal",
-			peerUserID:   userID,
-			storedUsers:  map[string]*types.User{},
-			idpEmail:     "idp@example.com",
-			idpHasData:   true,
-			expectEmail:  "idp@example.com",
-			expectUserID: userID,
-			expectIdPHit: true,
+			// The identity still resolves from the IdP, but an owner the store
+			// cannot resolve denies the fast-path rather than granting it.
+			name:               "idp email when stored user missing keeps peer.UserID as principal",
+			peerUserID:         userID,
+			storedUsers:        map[string]*types.User{},
+			idpEmail:           "idp@example.com",
+			idpHasData:         true,
+			expectEmail:        "idp@example.com",
+			expectUserID:       userID,
+			expectIdPHit:       true,
+			expectDeniedReason: deniedReasonUserNotFound,
 		},
 		{
 			name:         "unlinked peer uses peer name and never consults idp",
@@ -545,9 +631,13 @@ func TestValidateTunnelPeerUserEmailEnrichment(t *testing.T) {
 
 			require.NoError(t, err)
 			require.NotNil(t, resp)
-			assert.True(t, resp.GetValid(), "expected access granted")
+			assert.Equal(t, tt.expectDeniedReason == "", resp.GetValid(), "unexpected access decision")
+			assert.Equal(t, tt.expectDeniedReason, resp.GetDeniedReason(), "unexpected denied reason")
 			assert.Equal(t, tt.expectEmail, resp.GetUserEmail())
 			assert.Equal(t, tt.expectUserID, resp.GetUserId())
+			if tt.expectDeniedReason != "" {
+				assert.Empty(t, resp.GetSessionToken(), "a denied peer must not receive a session token")
+			}
 
 			if idpMock != nil {
 				if tt.expectIdPHit {
@@ -562,6 +652,193 @@ func TestValidateTunnelPeerUserEmailEnrichment(t *testing.T) {
 	}
 }
 
+// TestDeniedReasonValues pins the wire values of the account status denied
+// reasons. The proxy logs them and operators filter access logs on them, so a
+// rename is a breaking change rather than an internal detail.
+// TestSameAccount pins the fail-closed behaviour of the account binding: an
+// unset account on either side must never compare equal into a grant.
+func TestSameAccount(t *testing.T) {
+	assert.True(t, sameAccount("account1", "account1"), "matching accounts should bind")
+	assert.False(t, sameAccount("account1", "account2"), "different accounts must not bind")
+	assert.False(t, sameAccount("", ""), "two unset accounts must not bind")
+	assert.False(t, sameAccount("account1", ""), "an unset service account must not bind")
+	assert.False(t, sameAccount("", "account1"), "an unset user account must not bind")
+}
+
+func TestDeniedReasonValues(t *testing.T) {
+	assert.Equal(t, "pending_approval", deniedReasonPendingApproval, "pending approval denied reason wire value")
+	assert.Equal(t, "user_blocked", deniedReasonUserBlocked, "blocked user denied reason wire value")
+	assert.Equal(t, "user_not_found", deniedReasonUserNotFound, "unresolved user denied reason wire value")
+}
+
+// TestValidateTunnelPeerOwnerStatus verifies that the mesh fast-path gates on
+// the account status of the peer's owning user. A peer whose owner was blocked
+// after the peer registered must lose access, while an unlinked machine peer
+// keeps it.
+func TestValidateTunnelPeerOwnerStatus(t *testing.T) {
+	const (
+		domain    = "app.example.com"
+		accountID = "account1"
+		peerID    = "peer1"
+		peerName  = "peer-display-name"
+		userID    = "user1"
+	)
+
+	tests := []struct {
+		name               string
+		peerUserID         string
+		owner              *types.User
+		expectDeniedReason string
+		expectEmail        string
+	}{
+		{
+			name:       "active owner allowed",
+			peerUserID: userID,
+			owner:      &types.User{Id: userID, AccountID: accountID, Email: "user@example.com"},
+		},
+		{
+			name:               "owner pending approval denied",
+			peerUserID:         userID,
+			owner:              &types.User{Id: userID, AccountID: accountID, Email: "user@example.com", Blocked: true, PendingApproval: true},
+			expectDeniedReason: deniedReasonPendingApproval,
+		},
+		{
+			name:               "owner blocked after registering the peer denied",
+			peerUserID:         userID,
+			owner:              &types.User{Id: userID, AccountID: accountID, Email: "user@example.com", Blocked: true},
+			expectDeniedReason: deniedReasonUserBlocked,
+		},
+		{
+			name:       "unlinked machine peer stays allowed",
+			peerUserID: "",
+			owner:      &types.User{Id: userID, AccountID: accountID, Blocked: true},
+		},
+		{
+			// The user lookup is not account-scoped, so a peer row pointing at
+			// another account's user must not resolve into an owner: the peer is
+			// denied and the foreign email never reaches the response.
+			name:               "owner in another account denied and not disclosed",
+			peerUserID:         userID,
+			owner:              &types.User{Id: userID, AccountID: "otherAccount", Email: "foreign@example.com"},
+			expectDeniedReason: deniedReasonUserNotFound,
+			expectEmail:        peerName,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			svc := &service.Service{Domain: domain, AccountID: accountID}
+			usersManager := &mockUsersManager{users: map[string]*types.User{userID: tt.owner}}
+			server := &ProxyServiceServer{
+				serviceManager: &mockReverseProxyManager{
+					proxiesByAccount: map[string][]*service.Service{accountID: {svc}},
+				},
+				peersManager: &mockTunnelPeersManager{
+					peer: &peer.Peer{ID: peerID, Name: peerName, UserID: tt.peerUserID},
+				},
+				usersManager: usersManager,
+			}
+
+			resp, err := server.ValidateTunnelPeer(context.Background(), &proto.ValidateTunnelPeerRequest{
+				Domain:   domain,
+				TunnelIp: "100.64.0.1",
+			})
+
+			require.NoError(t, err)
+			require.NotNil(t, resp)
+			assert.Equal(t, tt.expectDeniedReason, resp.GetDeniedReason(), "unexpected denied reason")
+			assert.Equal(t, tt.expectDeniedReason == "", resp.GetValid(), "unexpected access decision")
+			if tt.expectDeniedReason != "" {
+				assert.Empty(t, resp.GetSessionToken(), "a denied peer must not receive a session token")
+			}
+
+			if tt.expectEmail != "" {
+				assert.Equal(t, tt.expectEmail, resp.GetUserEmail(), "unexpected identity on the response")
+			}
+
+			// The status gate and the identity resolution share one lookup;
+			// an unlinked peer has no owner to look up at all.
+			wantLookups := 1
+			if tt.peerUserID == "" {
+				wantLookups = 0
+			}
+			assert.Equal(t, wantLookups, usersManager.getUserCalls, "owner must be resolved exactly once per request")
+		})
+	}
+}
+
+// TestValidateTunnelPeerRecordsActivity pins that a granted mesh request is
+// handed to the activity manager. Which of those the manager then writes is its
+// own decision, covered by its tests.
+func TestValidateTunnelPeerRecordsActivity(t *testing.T) {
+	const (
+		domain    = "app.example.com"
+		accountID = "account1"
+		peerID    = "peer1"
+	)
+
+	activityManager := &mockActivityManager{}
+	server := &ProxyServiceServer{
+		activityManager: activityManager,
+		serviceManager: &mockReverseProxyManager{
+			proxiesByAccount: map[string][]*service.Service{
+				accountID: {{Domain: domain, AccountID: accountID}},
+			},
+		},
+		peersManager: &mockTunnelPeersManager{
+			peer: &peer.Peer{ID: peerID, Name: "agent", Status: &peer.PeerStatus{LastSeen: time.Now().Add(-3 * time.Hour)}},
+		},
+		usersManager: &mockUsersManager{users: map[string]*types.User{}},
+	}
+
+	resp, err := server.ValidateTunnelPeer(context.Background(), &proto.ValidateTunnelPeerRequest{
+		Domain:   domain,
+		TunnelIp: "100.64.0.1",
+	})
+
+	require.NoError(t, err)
+	require.True(t, resp.GetValid(), "peer should be granted access")
+
+	require.Len(t, activityManager.seenMarks, 1, "a granted peer should reach the activity manager once")
+	assert.Equal(t, accountID, activityManager.seenMarks[0].accountID, "activity must be attributed to the service account")
+	assert.Equal(t, peerID, activityManager.seenMarks[0].peerID, "activity must be attributed to the calling peer")
+}
+
+// TestValidateTunnelPeerDeniedRecordsNoActivity keeps the write on the granted
+// path only: a refused peer is not evidence its owner was active.
+func TestValidateTunnelPeerDeniedRecordsNoActivity(t *testing.T) {
+	const (
+		domain    = "app.example.com"
+		accountID = "account1"
+	)
+
+	activityManager := &mockActivityManager{}
+	server := &ProxyServiceServer{
+		activityManager: activityManager,
+		serviceManager: &mockReverseProxyManager{
+			proxiesByAccount: map[string][]*service.Service{
+				accountID: {{Domain: domain, AccountID: accountID}},
+			},
+		},
+		peersManager: &mockTunnelPeersManager{
+			peer: &peer.Peer{ID: "peer1", Name: "agent", UserID: "user1", Status: &peer.PeerStatus{LastSeen: time.Now().Add(-3 * time.Hour)}},
+		},
+		// The owner is blocked, so the tunnel gate denies before the mint.
+		usersManager: &mockUsersManager{users: map[string]*types.User{
+			"user1": {Id: "user1", AccountID: accountID, Blocked: true},
+		}},
+	}
+
+	resp, err := server.ValidateTunnelPeer(context.Background(), &proto.ValidateTunnelPeerRequest{
+		Domain:   domain,
+		TunnelIp: "100.64.0.1",
+	})
+
+	require.NoError(t, err)
+	require.False(t, resp.GetValid(), "blocked owner should be denied")
+	assert.Empty(t, activityManager.seenMarks, "a denied peer must not be marked seen")
+}
+
 func TestGetAccountProxyByDomain(t *testing.T) {
 	tests := []struct {
 		name             string

management/internals/shared/grpc/proxy_snapshot_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"google.golang.org/grpc"

management/internals/shared/grpc/server.go

@@ -1141,7 +1141,7 @@ func (s *Server) GetDeviceAuthorizationFlow(ctx context.Context, req *proto.Encr
 			Provider: proto.DeviceAuthorizationFlowProvider(provider),
 			ProviderConfig: &proto.ProviderConfig{
 				ClientID:           s.config.DeviceAuthorizationFlow.ProviderConfig.ClientID,
-				ClientSecret:       s.config.DeviceAuthorizationFlow.ProviderConfig.ClientSecret,
+				ClientSecret:       s.config.DeviceAuthorizationFlow.ProviderConfig.ClientSecret, //nolint:staticcheck
 				Domain:             s.config.DeviceAuthorizationFlow.ProviderConfig.Domain,
 				Audience:           s.config.DeviceAuthorizationFlow.ProviderConfig.Audience,
 				DeviceAuthEndpoint: s.config.DeviceAuthorizationFlow.ProviderConfig.DeviceAuthEndpoint,
@@ -1212,7 +1212,7 @@ func (s *Server) GetPKCEAuthorizationFlow(ctx context.Context, req *proto.Encryp
 			ProviderConfig: &proto.ProviderConfig{
 				Audience:              s.config.PKCEAuthorizationFlow.ProviderConfig.Audience,
 				ClientID:              s.config.PKCEAuthorizationFlow.ProviderConfig.ClientID,
-				ClientSecret:          s.config.PKCEAuthorizationFlow.ProviderConfig.ClientSecret,
+				ClientSecret:          s.config.PKCEAuthorizationFlow.ProviderConfig.ClientSecret, //nolint:staticcheck
 				TokenEndpoint:         s.config.PKCEAuthorizationFlow.ProviderConfig.TokenEndpoint,
 				AuthorizationEndpoint: s.config.PKCEAuthorizationFlow.ProviderConfig.AuthorizationEndpoint,
 				Scope:                 s.config.PKCEAuthorizationFlow.ProviderConfig.Scope,

management/internals/shared/grpc/sync_mappings_test.go

@@ -7,7 +7,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"google.golang.org/grpc"

management/internals/shared/grpc/token_mgr_test.go

@@ -10,7 +10,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/require"
 
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map"

management/internals/shared/grpc/validate_session_test.go

@@ -46,6 +46,7 @@ func setupValidateSessionTest(t *testing.T) *validateSessionTestSetup {
 	proxyService.SetServiceManager(serviceManager)
 
 	createTestProxies(t, ctx, testStore)
+	createStatusTestUsers(t, ctx, testStore)
 
 	return &validateSessionTestSetup{
 		proxyService: proxyService,
@@ -91,6 +92,82 @@ func createTestProxies(t *testing.T, ctx context.Context, testStore store.Store)
 		},
 	}
 	require.NoError(t, testStore.CreateService(ctx, restrictedProxy))
+
+	// Distributed to the account's "All" group, the configuration that hands a
+	// service to every user in the account.
+	allUsersProxy := &service.Service{
+		ID:                "allUsersProxyId",
+		AccountID:         "testAccountId",
+		Name:              "All Users Proxy",
+		Domain:            "all-users-proxy.example.com",
+		Enabled:           true,
+		SessionPrivateKey: privKey,
+		SessionPublicKey:  pubKey,
+		Auth: service.AuthConfig{
+			BearerAuth: &service.BearerAuthConfig{
+				Enabled:            true,
+				DistributionGroups: []string{allUsersGroupID},
+			},
+		},
+	}
+	require.NoError(t, testStore.CreateService(ctx, allUsersProxy))
+}
+
+const (
+	allUsersGroupID   = "allUsersGroupId"
+	pendingUserID     = "pendingUserId"
+	blockedUserID     = "blockedUserId"
+	pendingAllUsersID = "pendingAllUsersUserId"
+)
+
+// createStatusTestUsers adds the users whose account status must keep them out
+// of a proxy session. A user awaiting approval is persisted as both blocked and
+// pending approval, the way the approval flow stores one.
+func createStatusTestUsers(t *testing.T, ctx context.Context, testStore store.Store) {
+	t.Helper()
+
+	require.NoError(t, testStore.CreateGroup(ctx, &types.Group{
+		ID:        allUsersGroupID,
+		AccountID: "testAccountId",
+		Name:      "All",
+		Issued:    types.GroupIssuedAPI,
+	}))
+
+	users := []*types.User{
+		{
+			Id:              pendingUserID,
+			AccountID:       "testAccountId",
+			Role:            types.UserRoleUser,
+			AutoGroups:      []string{"allowedGroupId"},
+			Blocked:         true,
+			PendingApproval: true,
+			Issued:          "api",
+			CreatedAt:       time.Now(),
+		},
+		{
+			Id:              pendingAllUsersID,
+			AccountID:       "testAccountId",
+			Role:            types.UserRoleUser,
+			AutoGroups:      []string{allUsersGroupID},
+			Blocked:         true,
+			PendingApproval: true,
+			Issued:          "api",
+			CreatedAt:       time.Now(),
+		},
+		{
+			Id:              blockedUserID,
+			AccountID:       "testAccountId",
+			Role:            types.UserRoleUser,
+			AutoGroups:      []string{"allowedGroupId"},
+			Blocked:         true,
+			PendingApproval: false,
+			Issued:          "api",
+			CreatedAt:       time.Now(),
+		},
+	}
+	for _, user := range users {
+		require.NoError(t, testStore.SaveUser(ctx, user))
+	}
 }
 
 func generateSessionKeyPair(t *testing.T) (string, string) {
@@ -149,6 +226,114 @@ func TestValidateSession_UserNotInAllowedGroup(t *testing.T) {
 	assert.Empty(t, resp.GetPeerGroupIds(), "PeerGroupIds must mirror the resolved user's actual (empty) memberships on denial")
 }
 
+// TestValidateSession_PendingApprovalUserDenied covers a user who is a member of
+// the service's distribution group but is still waiting for an administrator to
+// approve the account. Group membership alone must not open the service.
+func TestValidateSession_PendingApprovalUserDenied(t *testing.T) {
+	setup := setupValidateSessionTest(t)
+	defer setup.cleanup()
+
+	proxy, err := setup.store.GetServiceByID(context.Background(), store.LockingStrengthNone, "testAccountId", "restrictedProxyId")
+	require.NoError(t, err)
+
+	token := createSessionToken(t, proxy.SessionPrivateKey, pendingUserID, "restricted-proxy.example.com")
+
+	resp, err := setup.proxyService.ValidateSession(context.Background(), &proto.ValidateSessionRequest{
+		Domain:       "restricted-proxy.example.com",
+		SessionToken: token,
+	})
+
+	require.NoError(t, err)
+	assert.False(t, resp.Valid, "User pending approval should be denied")
+	assert.Equal(t, deniedReasonPendingApproval, resp.DeniedReason, "Denied reason should name the pending approval state")
+	assert.Equal(t, pendingUserID, resp.UserId, "Denial should identify the user it applies to")
+	assert.Equal(t, []string{"allowedGroupId"}, resp.GetPeerGroupIds(), "PeerGroupIds must mirror the resolved user's group memberships on denial")
+	assert.Equal(t, []string{"Allowed Group"}, resp.GetPeerGroupNames(), "PeerGroupNames must pair with PeerGroupIds on denial")
+}
+
+// TestValidateSession_PendingApprovalUserInAllUsersGroupDenied covers the same
+// user against a service distributed to the account's "All" group, where every
+// user of the account is a member by default.
+func TestValidateSession_PendingApprovalUserInAllUsersGroupDenied(t *testing.T) {
+	setup := setupValidateSessionTest(t)
+	defer setup.cleanup()
+
+	proxy, err := setup.store.GetServiceByID(context.Background(), store.LockingStrengthNone, "testAccountId", "allUsersProxyId")
+	require.NoError(t, err)
+
+	token := createSessionToken(t, proxy.SessionPrivateKey, pendingAllUsersID, "all-users-proxy.example.com")
+
+	resp, err := setup.proxyService.ValidateSession(context.Background(), &proto.ValidateSessionRequest{
+		Domain:       "all-users-proxy.example.com",
+		SessionToken: token,
+	})
+
+	require.NoError(t, err)
+	assert.False(t, resp.Valid, "User pending approval should be denied even in the All Users group")
+	assert.Equal(t, deniedReasonPendingApproval, resp.DeniedReason, "Denied reason should name the pending approval state")
+	assert.Equal(t, pendingAllUsersID, resp.UserId, "Denial should identify the user it applies to")
+	assert.Equal(t, []string{allUsersGroupID}, resp.GetPeerGroupIds(), "PeerGroupIds must mirror the resolved user's group memberships on denial")
+}
+
+// TestValidateSession_BlockedUserDenied covers a user blocked after having been
+// approved, so PendingApproval is false and only the blocked flag is set.
+func TestValidateSession_BlockedUserDenied(t *testing.T) {
+	setup := setupValidateSessionTest(t)
+	defer setup.cleanup()
+
+	proxy, err := setup.store.GetServiceByID(context.Background(), store.LockingStrengthNone, "testAccountId", "restrictedProxyId")
+	require.NoError(t, err)
+
+	token := createSessionToken(t, proxy.SessionPrivateKey, blockedUserID, "restricted-proxy.example.com")
+
+	resp, err := setup.proxyService.ValidateSession(context.Background(), &proto.ValidateSessionRequest{
+		Domain:       "restricted-proxy.example.com",
+		SessionToken: token,
+	})
+
+	require.NoError(t, err)
+	assert.False(t, resp.Valid, "Blocked user should be denied")
+	assert.Equal(t, deniedReasonUserBlocked, resp.DeniedReason, "Denied reason should name the blocked state")
+	assert.Equal(t, blockedUserID, resp.UserId, "Denial should identify the user it applies to")
+}
+
+// TestValidateSession_UserAllowedAfterApproval walks the same session token
+// through the approval transition: denied while pending, allowed once an
+// administrator clears both flags.
+func TestValidateSession_UserAllowedAfterApproval(t *testing.T) {
+	setup := setupValidateSessionTest(t)
+	defer setup.cleanup()
+
+	ctx := context.Background()
+
+	proxy, err := setup.store.GetServiceByID(ctx, store.LockingStrengthNone, "testAccountId", "restrictedProxyId")
+	require.NoError(t, err)
+
+	token := createSessionToken(t, proxy.SessionPrivateKey, pendingUserID, "restricted-proxy.example.com")
+	req := &proto.ValidateSessionRequest{
+		Domain:       "restricted-proxy.example.com",
+		SessionToken: token,
+	}
+
+	resp, err := setup.proxyService.ValidateSession(ctx, req)
+	require.NoError(t, err)
+	require.False(t, resp.Valid, "User pending approval should be denied before approval")
+	assert.Equal(t, deniedReasonPendingApproval, resp.DeniedReason, "Denied reason should name the pending approval state")
+
+	user, err := setup.store.GetUserByUserID(ctx, store.LockingStrengthNone, pendingUserID)
+	require.NoError(t, err)
+	user.PendingApproval = false
+	user.Blocked = false
+	require.NoError(t, setup.store.SaveUser(ctx, user))
+
+	resp, err = setup.proxyService.ValidateSession(ctx, req)
+	require.NoError(t, err)
+	assert.True(t, resp.Valid, "Approved user should be allowed access")
+	assert.Empty(t, resp.DeniedReason)
+	assert.Equal(t, pendingUserID, resp.UserId, "Approved user should be identified in the response")
+	assert.Equal(t, []string{"allowedGroupId"}, resp.GetPeerGroupIds(), "PeerGroupIds must mirror the approved user's group memberships")
+}
+
 func TestValidateSession_UserInDifferentAccount(t *testing.T) {
 	setup := setupValidateSessionTest(t)
 	defer setup.cleanup()

management/server/account.go

@@ -33,6 +33,7 @@ import (
 	nbconfig "github.com/netbirdio/netbird/management/internals/server/config"
 	"github.com/netbirdio/netbird/management/server/account"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/affectedpeers"
 	nbcache "github.com/netbirdio/netbird/management/server/cache"
 	nbcontext "github.com/netbirdio/netbird/management/server/context"
 	"github.com/netbirdio/netbird/management/server/geolocation"
@@ -1626,6 +1627,8 @@ func (am *DefaultAccountManager) SyncUserJWTGroups(ctx context.Context, userAuth
 	var removeOldGroups []string
 	var hasChanges bool
 	var user *types.User
+	var change affectedpeers.Change
+	var snap *affectedpeers.Snapshot
 	err = am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		user, err = transaction.GetUserByUserID(ctx, store.LockingStrengthNone, userAuth.UserId)
 		if err != nil {
@@ -1664,14 +1667,25 @@ func (am *DefaultAccountManager) SyncUserJWTGroups(ctx context.Context, userAuth
 			return fmt.Errorf("error saving user: %w", err)
 		}
 
+		allGroupChanges := slices.Concat(addNewGroups, removeOldGroups)
+		// The user's auto-groups changed, so the SSH rules authorizing them ship a new
+		// group -> user mapping even when no peer moves between groups.
+		change.UserGroupIDs = allGroupChanges
+
+		// The user's peers are the changed entity in every scenario the sync can
+		// produce — group membership, IPv6 assignment, SSH mappings — so they refresh
+		// together with every peer they can connect to, like on a regular peer update.
+		userPeers, err := transaction.GetUserPeers(ctx, store.LockingStrengthNone, userAuth.AccountId, userAuth.UserId)
+		if err != nil {
+			return fmt.Errorf("error getting user peers: %w", err)
+		}
+		for _, peer := range userPeers {
+			change.ChangedPeerIDs = append(change.ChangedPeerIDs, peer.ID)
+		}
+
 		// Propagate changes to peers if group propagation is enabled
 		if settings.GroupsPropagationEnabled {
-			peers, err := transaction.GetUserPeers(ctx, store.LockingStrengthNone, userAuth.AccountId, userAuth.UserId)
-			if err != nil {
-				return fmt.Errorf("error getting user peers: %w", err)
-			}
-
-			for _, peer := range peers {
+			for _, peer := range userPeers {
 				for _, g := range addNewGroups {
 					if err := transaction.AddPeerToGroup(ctx, userAuth.AccountId, peer.ID, g); err != nil {
 						return fmt.Errorf("error adding peer %s to group %s: %w", peer.ID, g, err)
@@ -1684,7 +1698,8 @@ func (am *DefaultAccountManager) SyncUserJWTGroups(ctx context.Context, userAuth
 				}
 			}
 
-			allGroupChanges := slices.Concat(addNewGroups, removeOldGroups)
+			change.LinkGroups = allGroupChanges
+
 			if err = am.reconcileIPv6ForGroupChanges(ctx, transaction, userAuth.AccountId, allGroupChanges); err != nil {
 				return fmt.Errorf("reconcile IPv6 for group changes: %w", err)
 			}
@@ -1694,6 +1709,10 @@ func (am *DefaultAccountManager) SyncUserJWTGroups(ctx context.Context, userAuth
 			}
 		}
 
+		if snap, err = affectedpeers.Load(ctx, transaction, userAuth.AccountId, change); err != nil {
+			return err
+		}
+
 		return nil
 	})
 	if err != nil {
@@ -1730,20 +1749,17 @@ func (am *DefaultAccountManager) SyncUserJWTGroups(ctx context.Context, userAuth
 		}
 	}
 
-	removedGroupAffectsPeers, err := areGroupChangesAffectPeers(ctx, am.Store, userAuth.AccountId, removeOldGroups)
-	if err != nil {
-		return err
-	}
-
-	newGroupsAffectsPeers, err := areGroupChangesAffectPeers(ctx, am.Store, userAuth.AccountId, addNewGroups)
-	if err != nil {
-		return err
-	}
-
-	if removedGroupAffectsPeers || newGroupsAffectsPeers {
-		log.WithContext(ctx).Tracef("user %s: JWT group membership changed, updating account peers", userAuth.UserId)
-		am.BufferUpdateAccountPeers(ctx, userAuth.AccountId, types.UpdateReason{Resource: types.UpdateResourceUser, Operation: types.UpdateOperationUpdate})
-	}
+	log.WithContext(ctx).Tracef("user %s: JWT group membership changed, updating affected peers", userAuth.UserId)
+	bgCtx := context.WithoutCancel(ctx)
+	go func() {
+		affectedPeerIDs := snap.Expand(bgCtx, userAuth.AccountId, change)
+		if len(affectedPeerIDs) == 0 {
+			return
+		}
+		if err := am.networkMapController.BufferUpdateAffectedPeers(bgCtx, userAuth.AccountId, affectedPeerIDs, types.UpdateReason{Resource: types.UpdateResourceUser, Operation: types.UpdateOperationUpdate}); err != nil {
+			log.WithContext(bgCtx).Errorf("failed to update affected peers after JWT group sync for account %s: %v", userAuth.AccountId, err)
+		}
+	}()
 
 	return nil
 }
@@ -2426,30 +2442,24 @@ func (am *DefaultAccountManager) reconcileIPv6ForGroupChanges(ctx context.Contex
 		return fmt.Errorf("get account settings: %w", err)
 	}
 
-	if len(settings.IPv6EnabledGroups) == 0 {
-		return nil
-	}
-
-	enabledSet := make(map[string]struct{}, len(settings.IPv6EnabledGroups))
-	for _, gid := range settings.IPv6EnabledGroups {
-		enabledSet[gid] = struct{}{}
-	}
-
-	affected := false
-	for _, gid := range groupIDs {
-		if _, ok := enabledSet[gid]; ok {
-			affected = true
-			break
-		}
-	}
-
-	if !affected {
+	if !ipv6ReconcileNeeded(settings, groupIDs) {
 		return nil
 	}
 
 	return am.updatePeerIPv6Addresses(ctx, transaction, accountID, settings)
 }
 
+// ipv6ReconcileNeeded reports whether changes to the given groups trigger an IPv6
+// reconciliation.
+func ipv6ReconcileNeeded(settings *types.Settings, groupIDs []string) bool {
+	for _, groupID := range groupIDs {
+		if slices.Contains(settings.IPv6EnabledGroups, groupID) {
+			return true
+		}
+	}
+	return false
+}
+
 func (am *DefaultAccountManager) ensureIPv6Subnet(ctx context.Context, transaction store.Store, accountID string, settings *types.Settings, network *types.Network) error {
 	if settings.NetworkRangeV6.IsValid() {
 		network.NetV6 = net.IPNet{

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"

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_test.go

@@ -14,7 +14,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/prometheus/client_golang/prometheus/push"
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
@@ -1757,6 +1757,7 @@ func TestAccount_Copy(t *testing.T) {
 				AccountID: "account1",
 			},
 		},
+		PostureValidation: map[string]map[string]bool{"1": {"1": true}},
 	}
 	err := hasNilField(account)
 	if err != nil {

management/server/activity/codes.go

@@ -281,6 +281,9 @@ const (
 	// 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
 )
 
@@ -453,6 +456,7 @@ var activityMap = map[Activity]Code{
 	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"},

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_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/resolver.go

@@ -18,6 +18,7 @@ import (
 	"context"
 
 	log "github.com/sirupsen/logrus"
+	"golang.org/x/exp/maps"
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
@@ -83,7 +84,7 @@ func (snap *Snapshot) loadCollections(ctx context.Context, s store.Store, accoun
 	hasGroupOrPeerChange := len(c.ChangedGroupIDs) > 0 || len(c.ChangedPeerIDs) > 0 || len(c.LinkGroups) > 0 || len(c.Resources) > 0
 	hasNetworkObject := len(c.Routers) > 0 || len(c.Resources) > 0 || len(c.Networks) > 0
 	// the resource<->router bridge can fire for any of these
-	needsRoutersResources := hasGroupOrPeerChange || len(c.PostureCheckIDs) > 0 || len(c.Policies) > 0 || hasNetworkObject
+	needsRoutersResources := hasGroupOrPeerChange || len(c.PostureCheckIDs) > 0 || len(c.Policies) > 0 || hasNetworkObject || len(c.UserGroupIDs) > 0 || c.AllowedUsersChanged
 
 	if needsRoutersResources {
 		if err := snap.loadPolicyRoutersResources(ctx, s, accountID); err != nil {
@@ -219,6 +220,18 @@ type Change struct {
 	// (correct when the peer's own attributes changed, e.g. IP/status).
 	OutputPeerIDs []string
 
+	// UserGroupIDs are groups whose USER membership changed (a user's auto-groups),
+	// as opposed to their peer membership. Peers ship the group -> user mapping only
+	// for the groups an SSH rule authorizes, so these refresh the destinations of the
+	// SSH rules authorizing them — independently of any peer moving between groups.
+	UserGroupIDs []string
+
+	// AllowedUsersChanged marks a change to the set of users allowed to open SSH
+	// sessions — a user was created, blocked or unblocked. That set is account-wide,
+	// and peers receive it through the SSH rules that name no group or user of their
+	// own, so those rules' destinations refresh.
+	AllowedUsersChanged bool
+
 	// LinkGroups are groups used ONLY to match policies/routes/routers and walk to the
 	// OPPOSITE side — they are never expanded to their own members. Use this when a
 	// peer's group membership changed: pass the peer in ChangedPeerIDs and its
@@ -240,6 +253,8 @@ func (c Change) isEmpty() bool {
 		len(c.Resources) == 0 &&
 		len(c.Networks) == 0 &&
 		len(c.PostureCheckIDs) == 0 &&
+		len(c.UserGroupIDs) == 0 &&
+		!c.AllowedUsersChanged &&
 		len(c.DistributionGroupIDs) == 0 &&
 		len(c.RemovedPeersByGroup) == 0 &&
 		len(c.LinkGroups) == 0 &&
@@ -359,6 +374,9 @@ func (r *resolver) walk() {
 		r.collectFromProxyServices()
 	}
 
+	r.collectFromSSHAuthorizedGroups()
+	r.collectFromAllowedUsers()
+
 	r.collectFromChangedRoutes(r.change.Routes)
 	r.collectFromChangedRouters(r.change.Routers)
 	r.collectFromChangedResources(r.change.Resources)
@@ -811,6 +829,59 @@ func (r *resolver) collectFromNameServers() {
 	}
 }
 
+// collectFromSSHAuthorizedGroups folds the destinations of the enabled SSH rules that
+// authorize a group whose user membership changed. Those destination peers carry the
+// group -> user mapping for the groups they authorize, so they refresh even when no
+// peer moved between groups.
+func (r *resolver) collectFromSSHAuthorizedGroups() {
+	if len(r.change.UserGroupIDs) == 0 {
+		return
+	}
+
+	changed := toSet(r.change.UserGroupIDs)
+	for _, policy := range r.policies() {
+		for _, rule := range policy.Rules {
+			if !rule.Enabled || rule.Protocol != types.PolicyRuleProtocolNetbirdSSH {
+				continue
+			}
+			if !anyInSet(maps.Keys(rule.AuthorizedGroups), changed) {
+				continue
+			}
+			log.WithContext(r.ctx).Tracef("collectFromSSHAuthorizedGroups: rule %s authorizes a changed user group -> folding its destinations", rule.ID)
+			r.foldPolicySideForRule(policy, rule, sideDestination)
+		}
+	}
+}
+
+// collectFromAllowedUsers folds the destinations of the rules that make a peer carry
+// the account's allowed-user set, for a change to who is in that set.
+func (r *resolver) collectFromAllowedUsers() {
+	if !r.change.AllowedUsersChanged {
+		return
+	}
+
+	for _, policy := range r.policies() {
+		for _, rule := range policy.Rules {
+			if !rule.Enabled || !ruleShipsAllowedUsers(rule) {
+				continue
+			}
+			log.WithContext(r.ctx).Tracef("collectFromAllowedUsers: rule %s ships the allowed-user set -> folding its destinations", rule.ID)
+			r.foldPolicySideForRule(policy, rule, sideDestination)
+		}
+	}
+}
+
+// ruleShipsAllowedUsers reports whether a rule makes its destination peers carry the
+// account's allowed-user set. It mirrors the network map's SSH requirements except for
+// the destination peer's own SSH flag, which the snapshot does not hold — so it folds a
+// superset and never misses a peer.
+func ruleShipsAllowedUsers(rule *types.PolicyRule) bool {
+	if rule.Protocol == types.PolicyRuleProtocolNetbirdSSH {
+		return len(rule.AuthorizedGroups) == 0 && rule.AuthorizedUser == ""
+	}
+	return types.PolicyRuleImpliesLegacySSH(rule)
+}
+
 func (r *resolver) collectFromDNSSettings() {
 	if len(r.linkGroups) == 0 || r.snap.dnsSettings == nil {
 		return

management/server/affectedpeers/resolver_test.go

@@ -85,6 +85,8 @@ func TestChangeIsEmpty(t *testing.T) {
 	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) {

management/server/agentnetwork_budgetrule_realstack_test.go

@@ -68,7 +68,10 @@ func TestAgentNetwork_BudgetRuleCRUD_RealManager(t *testing.T) {
 
 // TestAgentNetwork_UpdateSettings_PreservesImmutableAndTogglesCollection is the
 // GC-1 guard for UpdateSettings: it must apply the collection toggles while
-// preserving the immutable Cluster/Subdomain pinned at bootstrap.
+// 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")
@@ -84,7 +87,14 @@ func TestAgentNetwork_UpdateSettings_PreservesImmutableAndTogglesCollection(t *t
 
 	mgr := agentnetwork.NewManager(am.Store, permissions.NewManager(am.Store), am, nil)
 
-	// Creating a provider bootstraps the settings row (cluster + subdomain).
+	// 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",
@@ -93,43 +103,64 @@ func TestAgentNetwork_UpdateSettings_PreservesImmutableAndTogglesCollection(t *t
 		APIKey:      "sk-test",
 		Enabled:     true,
 		Models:      []agenttypes.ProviderModel{{ID: "gpt-5.4"}},
-	}, clusterAddr)
-	require.NoError(t, err, "CreateProvider must bootstrap settings")
-
-	before, err := mgr.GetSettings(ctx, accountID, adminUserID)
-	require.NoError(t, err, "GetSettings must succeed after bootstrap")
-	require.Equal(t, clusterAddr, before.Cluster, "cluster pinned at bootstrap")
-	require.NotEmpty(t, before.Subdomain, "subdomain pinned at bootstrap")
-	assert.False(t, before.EnablePromptCollection, "prompt collection defaults off")
-
-	// A cluster different from the one pinned at bootstrap must be rejected
-	// outright — never silently swapped or ignored.
-	_, err = mgr.UpdateSettings(ctx, adminUserID, &agenttypes.Settings{
-		AccountID:           accountID,
-		Cluster:             "attacker.cluster",
-		EnableLogCollection: true,
 	})
-	require.Error(t, err, "UpdateSettings with a mismatched cluster must fail")
+	require.NoError(t, err, "CreateProvider must succeed")
 
-	// Flipping the toggles works with the pinned cluster echoed back (and
-	// with it omitted); the subdomain is never taken from the request.
+	// 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,
-		Cluster:                clusterAddr,
-		Subdomain:              "evil",
+		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.Cluster, updated.Cluster, "cluster is immutable and must be preserved")
-	assert.Equal(t, before.Subdomain, updated.Subdomain, "subdomain is immutable and must be preserved")
+	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.Cluster, reloaded.Cluster, "persisted cluster unchanged")
+	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

@@ -92,6 +92,14 @@ func TestAgentNetwork_ProviderCRUD_FansOutToProxyAndClientPeers(t *testing.T) {
 	// 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",
@@ -100,7 +108,7 @@ func TestAgentNetwork_ProviderCRUD_FansOutToProxyAndClientPeers(t *testing.T) {
 		APIKey:      "sk-test-key",
 		Enabled:     true,
 		Models:      []agenttypes.ProviderModel{{ID: "gpt-5.4"}},
-	}, clusterAddr)
+	})
 	require.NoError(t, err, "CreateProvider must succeed")
 
 	policy, err := agentMgr.CreatePolicy(ctx, adminUserID, &agenttypes.Policy{

management/server/dns_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 
 	nbdns "github.com/netbirdio/netbird/dns"

management/server/group_test.go

@@ -11,7 +11,7 @@ import (
 	"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"

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

@@ -12,7 +12,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 

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

@@ -10,7 +10,7 @@ import (
 	"net/mail"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/gorilla/mux"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"

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().

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"
 

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

@@ -2,6 +2,7 @@ package proxy
 
 import (
 	"context"
+	"errors"
 	"net"
 	"net/http"
 	"net/netip"
@@ -108,7 +109,7 @@ func (h *AuthCallbackHandler) handleCallback(w http.ResponseWriter, r *http.Requ
 		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 +125,20 @@ 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"
+	}
+	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"
@@ -221,6 +222,7 @@ func setupAuthCallbackTest(t *testing.T) *testSetup {
 	)
 
 	proxyService.SetServiceManager(&testServiceManager{store: testStore})
+	proxyService.SetActivityManager(activitymanager.NewManager(testStore))
 
 	handler := NewAuthCallbackHandler(proxyService, nil)
 
@@ -360,6 +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.
@@ -490,6 +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

@@ -64,6 +64,19 @@ func (h *handler) createSetupKey(w http.ResponseWriter, r *http.Request) {
 		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

@@ -134,6 +134,40 @@ 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/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/identity_provider_test.go

@@ -10,7 +10,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

management/server/instance/setup_service_test.go

@@ -5,7 +5,7 @@ import (
 	"errors"
 	"testing"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

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"

management/server/management_test.go

@@ -10,7 +10,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	pb "github.com/golang/protobuf/proto" //nolint
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"

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

@@ -736,3 +736,125 @@ func TestFoldCostAggregatesIntoBuckets_SkipsAlreadyMigrated(t *testing.T) {
 	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/nameserver_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 

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/peer_test.go

@@ -16,7 +16,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
+	"go.uber.org/mock/gomock"
 	"github.com/rs/xid"
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"

management/server/permissions/manager.go

@@ -1,6 +1,6 @@
 package permissions
 
-//go:generate go run github.com/golang/mock/mockgen -package permissions -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//go:generate go tool mockgen -package permissions -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
 
 import (
 	"context"

management/server/permissions/manager_mock.go

@@ -1,5 +1,10 @@
 // Code generated by MockGen. DO NOT EDIT.
 // Source: ./manager.go
+//
+// Generated by this command:
+//
+//	mockgen -package permissions -destination=manager_mock.go -source=./manager.go -build_flags=-mod=mod
+//
 
 // Package permissions is a generated GoMock package.
 package permissions
@@ -8,18 +13,19 @@ import (
 	context "context"
 	reflect "reflect"
 
-	gomock "github.com/golang/mock/gomock"
 	account "github.com/netbirdio/netbird/management/server/account"
 	modules "github.com/netbirdio/netbird/management/server/permissions/modules"
 	operations "github.com/netbirdio/netbird/management/server/permissions/operations"
 	roles "github.com/netbirdio/netbird/management/server/permissions/roles"
 	types "github.com/netbirdio/netbird/management/server/types"
+	gomock "go.uber.org/mock/gomock"
 )
 
 // MockManager is a mock of Manager interface.
 type MockManager struct {
 	ctrl     *gomock.Controller
 	recorder *MockManagerMockRecorder
+	isgomock struct{}
 }
 
 // MockManagerMockRecorder is the mock recorder for MockManager.
@@ -49,7 +55,7 @@ func (m *MockManager) GetPermissionsByRole(ctx context.Context, role types.UserR
 }
 
 // GetPermissionsByRole indicates an expected call of GetPermissionsByRole.
-func (mr *MockManagerMockRecorder) GetPermissionsByRole(ctx, role interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) GetPermissionsByRole(ctx, role any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetPermissionsByRole", reflect.TypeOf((*MockManager)(nil).GetPermissionsByRole), ctx, role)
 }
@@ -61,7 +67,7 @@ func (m *MockManager) SetAccountManager(accountManager account.Manager) {
 }
 
 // SetAccountManager indicates an expected call of SetAccountManager.
-func (mr *MockManagerMockRecorder) SetAccountManager(accountManager interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SetAccountManager(accountManager any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetAccountManager", reflect.TypeOf((*MockManager)(nil).SetAccountManager), accountManager)
 }
@@ -76,7 +82,7 @@ func (m *MockManager) ValidateAccountAccess(ctx context.Context, accountID strin
 }
 
 // ValidateAccountAccess indicates an expected call of ValidateAccountAccess.
-func (mr *MockManagerMockRecorder) ValidateAccountAccess(ctx, accountID, user, allowOwnerAndAdmin interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ValidateAccountAccess(ctx, accountID, user, allowOwnerAndAdmin any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ValidateAccountAccess", reflect.TypeOf((*MockManager)(nil).ValidateAccountAccess), ctx, accountID, user, allowOwnerAndAdmin)
 }
@@ -90,7 +96,7 @@ func (m *MockManager) ValidateRoleModuleAccess(ctx context.Context, accountID st
 }
 
 // ValidateRoleModuleAccess indicates an expected call of ValidateRoleModuleAccess.
-func (mr *MockManagerMockRecorder) ValidateRoleModuleAccess(ctx, accountID, role, module, operation interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ValidateRoleModuleAccess(ctx, accountID, role, module, operation any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ValidateRoleModuleAccess", reflect.TypeOf((*MockManager)(nil).ValidateRoleModuleAccess), ctx, accountID, role, module, operation)
 }
@@ -106,7 +112,7 @@ func (m *MockManager) ValidateUserPermissions(ctx context.Context, accountID, us
 }
 
 // ValidateUserPermissions indicates an expected call of ValidateUserPermissions.
-func (mr *MockManagerMockRecorder) ValidateUserPermissions(ctx, accountID, userID, module, operation interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) ValidateUserPermissions(ctx, accountID, userID, module, operation any) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ValidateUserPermissions", reflect.TypeOf((*MockManager)(nil).ValidateUserPermissions), ctx, accountID, userID, module, operation)
 }