Merge branch 'main' into fix/pkce-flow-session-extend

Both sides extended the OAuth flow entry points: main threads a login hint
through GetOAuthFlow/NewOAuthFlow, this branch threads a sessionExtend flag.
The merged signatures carry both.

client/android/login.go keeps main's structure — the hint reaches the flow
through GetOAuthFlow rather than a loginHintSetter assertion, profile email
helpers live in client/mobile, and runOAuthFlow is the shared driver — with
this branch's sessionExtend plumbing and wrong-account retry layered on top.
The branch's runInteractiveFlow is dropped in favour of runOAuthFlow.

client/android/ssh_client.go is new from main and merged without conflict,
but calls NewOAuthFlow; SSH login is a fresh login, so it passes false.

management.pb.go regenerated with the pinned protoc-gen-go v1.26 from
generate.sh. The .proto merged cleanly: main's LazyState and proxy_embedded
are disjoint from this branch's SessionExtend field.
This commit is contained in:
Zoltán Papp
2026-08-26 16:00:34 +02:00
346 changed files with 20217 additions and 4994 deletions
+2 -2
View File
@@ -8,7 +8,7 @@ import (
"testing"
"time"
"github.com/golang/mock/gomock"
"go.uber.org/mock/gomock"
log "github.com/sirupsen/logrus"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
@@ -591,7 +591,7 @@ func Test_GetPKCEAuthorizationFlow(t *testing.T) {
expectedFlowInfo := &mgmtProto.PKCEAuthorizationFlow{
ProviderConfig: &mgmtProto.ProviderConfig{
ClientID: "client",
ClientSecret: "secret",
ClientSecret: "secret", //nolint:staticcheck
},
}
+57 -18
View File
@@ -21,6 +21,7 @@ import (
"google.golang.org/grpc/connectivity"
nbgrpc "github.com/netbirdio/netbird/client/grpc"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/system"
"github.com/netbirdio/netbird/encryption"
"github.com/netbirdio/netbird/shared/management/domain"
@@ -62,6 +63,10 @@ type GrpcClient struct {
connStateCallbackLock sync.RWMutex
serverURL string
// netMgr gates the stream retry loop on OS-reported network
// availability and sweeps the transport on network change.
netMgr *netevents.Manager
// syncStreamErr holds the last Sync stream error, or nil while the stream
// is established and healthy. GetServerKey succeeds even when the peer
// cannot sync (e.g. the server returns "settings not found"), so the
@@ -111,16 +116,37 @@ func MaxRecvMsgSize() int {
return size
}
// Option configures optional GrpcClient behavior.
type Option func(*GrpcClient)
// WithNetEvents injects the OS network event handling.
func WithNetEvents(events *netevents.Manager) Option {
return func(c *GrpcClient) { c.netMgr = events }
}
// NewClient creates a new client to Management service
func NewClient(ctx context.Context, addr string, ourPrivateKey wgtypes.Key, tlsEnabled bool) (*GrpcClient, error) {
var conn *grpc.ClientConn
func NewClient(ctx context.Context, addr string, ourPrivateKey wgtypes.Key, tlsEnabled bool, opts ...Option) (*GrpcClient, error) {
// Options apply before dialing: the sweeper must wrap the first connection too.
c := &GrpcClient{
key: ourPrivateKey,
ctx: ctx,
connStateCallbackLock: sync.RWMutex{},
serverURL: addr,
}
for _, opt := range opts {
opt(c)
}
var extraOpts []grpc.DialOption
if maxSize := MaxRecvMsgSize(); maxSize > 0 {
extraOpts = append(extraOpts, grpc.WithDefaultCallOptions(grpc.MaxCallRecvMsgSize(maxSize)))
log.Infof("management gRPC max receive message size set to %d bytes", maxSize)
}
if c.netMgr != nil {
extraOpts = append(extraOpts, nbgrpc.WithSweeper(c.netMgr))
}
var conn *grpc.ClientConn
operation := func() error {
var err error
conn, err = nbgrpc.CreateConnection(ctx, addr, tlsEnabled, wsproxy.ManagementComponent, extraOpts...)
@@ -136,16 +162,9 @@ func NewClient(ctx context.Context, addr string, ourPrivateKey wgtypes.Key, tlsE
return nil, err
}
realClient := proto.NewManagementServiceClient(conn)
return &GrpcClient{
key: ourPrivateKey,
realClient: realClient,
ctx: ctx,
conn: conn,
connStateCallbackLock: sync.RWMutex{},
serverURL: addr,
}, nil
c.conn = conn
c.realClient = proto.NewManagementServiceClient(conn)
return c, nil
}
// GetServerURL returns the management server URL
@@ -206,16 +225,36 @@ func (c *GrpcClient) withMgmtStream(
ctx context.Context,
handler func(ctx context.Context, serverPubKey wgtypes.Key, backOff backoff.BackOff) error,
) error {
backOff := defaultBackoff(ctx)
backOff := c.netMgr.QuickRetryBackoff(ctx, defaultBackoff(ctx))
operation := func() error {
log.Debugf("management connection state %v", c.conn.GetState())
connState := c.conn.GetState()
// suspend reconnect attempts while the OS reports no usable network.
// Wait only errors on a cancelled context, which means shutdown, so
// stop the loop without reporting a failure.
if waited, err := c.netMgr.Wait(ctx); err != nil {
log.Debugf("management connection context has been canceled while offline, this usually indicates shutdown")
return nil //nolint:nilerr // a cancelled context means shutdown, not a retryable failure
} else if waited {
backOff.Reset()
// dials attempted while offline grew the channel's internal backoff;
// reset it too, or the reconnect waits out that timer first
c.conn.ResetConnectBackoff()
}
connState := c.conn.GetState()
log.Debugf("management connection state %v", connState)
if connState == connectivity.Shutdown {
return backoff.Permanent(fmt.Errorf("connection to management has been shut down"))
} else if !(connState == connectivity.Ready || connState == connectivity.Idle) {
}
if !(connState == connectivity.Ready || connState == connectivity.Idle) {
// A dial may already be in flight (e.g. the other stream triggered
// it after a network change); wait for it to settle and proceed if
// the channel became usable, instead of burning a backoff round on
// a successful dial. A failed dial errors out as before.
c.conn.WaitForStateChange(ctx, connState)
return fmt.Errorf("connection to management is not ready and in %s state", connState)
connState = c.conn.GetState()
if !(connState == connectivity.Ready || connState == connectivity.Idle) {
return fmt.Errorf("connection to management is not ready and in %s state", connState)
}
}
serverPubKey, err := c.getServerPublicKey()
@@ -227,7 +266,7 @@ func (c *GrpcClient) withMgmtStream(
return handler(ctx, *serverPubKey, backOff)
}
err := backoff.Retry(operation, backOff)
err := nbgrpc.Retry(ctx, operation, backOff, c.netMgr)
if err != nil {
log.Warnf("exiting the Management service connection retry loop due to the unrecoverable error: %s", err)
}
+115 -1
View File
@@ -4608,7 +4608,7 @@ components:
FleetDMMatchAttributes:
type: object
description: Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open-source version. Premium-only attributes are marked accordingly
description: Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open source version. Premium-only attributes are marked accordingly
additionalProperties: false
properties:
disk_encryption_enabled:
@@ -5335,6 +5335,84 @@ components:
- input_per_1k
- output_per_1k
- context_window
AgentNetworkModelDiscoveryRequest:
type: object
properties:
catalog_provider_id:
type: string
description: Catalog provider to query (AgentNetworkCatalogProvider.id). Determines the listing endpoint, the auth header and the response shape.
example: "bedrock_api"
upstream_url:
type: string
description: |
The upstream being configured. Used to reach vendors that serve their listing from the same host as inference, and to read back the region for those whose host embeds one. Ignored when provider_id is supplied.
example: "https://bedrock-runtime.eu-central-1.amazonaws.com"
api_key:
type: string
description: Credential to query the vendor with, for a provider that has not been saved yet. Mutually exclusive with provider_id.
example: "sk-..."
provider_id:
type: string
description: Existing Agent Network provider record whose stored credential and upstream should be used. Lets the form refresh the list without the client holding the key.
example: "ch8i4ug6lnn4g9hqv7m0"
required:
- catalog_provider_id
AgentNetworkModelDiscoveryResponse:
type: object
properties:
models:
type: array
description: Models the credential can reach, in the order the vendor returned them.
items:
$ref: '#/components/schemas/AgentNetworkDiscoveredModel'
required:
- models
AgentNetworkDiscoveredModel:
type: object
properties:
id:
type: string
description: |
Identifier to register on the provider record, in the form the vendor issues it. For Bedrock this is the region-prefixed inference-profile id, which is the only form AWS accepts at invoke time.
example: "eu.anthropic.claude-haiku-4-5-20251001-v1:0"
label:
type: string
description: Vendor-supplied display name, where the vendor supplies one.
example: "EU Anthropic Claude Haiku 4.5"
pricing_known:
type: boolean
description: Whether NetBird's shipped pricing table can price this model. When false the rates below are all zero and the operator must set them, or requests to this model would record a cost of zero.
example: true
input_per_1k:
type: number
format: double
description: Default input token price per 1k tokens, in USD, from the same table the proxy bills with. Zero when pricing_known is false.
example: 0.005
output_per_1k:
type: number
format: double
description: Default output token price per 1k tokens, in USD. Zero when pricing_known is false.
example: 0.015
cached_input_per_1k:
type: number
format: double
description: OpenAI-shape cache rate — default cost per 1k cached prompt tokens (a subset of input tokens), in USD. Absent when the model has no cached-input discount.
example: 0.000075
cache_read_per_1k:
type: number
format: double
description: Anthropic-shape cache rate — default cost per 1k cache-read tokens (additive to input tokens), in USD. Absent when the model has no cache-read rate.
example: 0.0003
cache_creation_per_1k:
type: number
format: double
description: Anthropic-shape cache rate — default cost per 1k cache-creation tokens (additive to input tokens), in USD. Absent when the model has no cache-creation rate.
example: 0.00375
required:
- id
- pricing_known
- input_per_1k
- output_per_1k
AgentNetworkCatalogProvider:
type: object
properties:
@@ -14004,6 +14082,42 @@ paths:
"$ref": "#/components/responses/forbidden"
'500':
"$ref": "#/components/responses/internal_error"
/api/agent-network/catalog/providers/models:
post:
summary: Discover the models a provider credential can reach
description: |
Asks the vendor which models the supplied credential can actually use, so the provider form can offer a live list instead of only the static catalog. The endpoint, auth header and response shape are taken from the catalog entry, never from the request.
Supply either an api_key together with the upstream_url being configured (before the provider is saved), or a provider_id of an existing record to reuse its stored credential.
Returns 422 for a catalog provider that has no listing endpoint (most gateways); the caller should fall back to the catalog's own model list. A model whose price the shipped table does not know is returned with pricing_known false, and the operator must set rates for it.
tags: [ Agent Network ]
security:
- BearerAuth: [ ]
- TokenAuth: [ ]
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/AgentNetworkModelDiscoveryRequest'
responses:
'200':
description: The models the credential can reach
content:
application/json:
schema:
$ref: '#/components/schemas/AgentNetworkModelDiscoveryResponse'
'400':
"$ref": "#/components/responses/bad_request"
'401':
"$ref": "#/components/responses/requires_authentication"
'403':
"$ref": "#/components/responses/forbidden"
'422':
"$ref": "#/components/responses/validation_failed_simple"
'500':
"$ref": "#/components/responses/internal_error"
/api/agent-network/providers:
get:
summary: List all Agent Network Providers
+54 -3
View File
@@ -2120,6 +2120,33 @@ type AgentNetworkConsumption struct {
// AgentNetworkConsumptionDimensionKind Whether this row counts a single end user or a single source group across every member.
type AgentNetworkConsumptionDimensionKind string
// AgentNetworkDiscoveredModel defines model for AgentNetworkDiscoveredModel.
type AgentNetworkDiscoveredModel struct {
// CacheCreationPer1k Anthropic-shape cache rate — default cost per 1k cache-creation tokens (additive to input tokens), in USD. Absent when the model has no cache-creation rate.
CacheCreationPer1k *float64 `json:"cache_creation_per_1k,omitempty"`
// CacheReadPer1k Anthropic-shape cache rate — default cost per 1k cache-read tokens (additive to input tokens), in USD. Absent when the model has no cache-read rate.
CacheReadPer1k *float64 `json:"cache_read_per_1k,omitempty"`
// CachedInputPer1k OpenAI-shape cache rate — default cost per 1k cached prompt tokens (a subset of input tokens), in USD. Absent when the model has no cached-input discount.
CachedInputPer1k *float64 `json:"cached_input_per_1k,omitempty"`
// Id Identifier to register on the provider record, in the form the vendor issues it. For Bedrock this is the region-prefixed inference-profile id, which is the only form AWS accepts at invoke time.
Id string `json:"id"`
// InputPer1k Default input token price per 1k tokens, in USD, from the same table the proxy bills with. Zero when pricing_known is false.
InputPer1k float64 `json:"input_per_1k"`
// Label Vendor-supplied display name, where the vendor supplies one.
Label *string `json:"label,omitempty"`
// OutputPer1k Default output token price per 1k tokens, in USD. Zero when pricing_known is false.
OutputPer1k float64 `json:"output_per_1k"`
// PricingKnown Whether NetBird's shipped pricing table can price this model. When false the rates below are all zero and the operator must set them, or requests to this model would record a cost of zero.
PricingKnown bool `json:"pricing_known"`
}
// AgentNetworkGuardrail defines model for AgentNetworkGuardrail.
type AgentNetworkGuardrail struct {
// Checks Guardrail check parameters. Each entry has an `enabled` flag plus per-check configuration; disabled entries are inert.
@@ -2167,6 +2194,27 @@ type AgentNetworkGuardrailRequest struct {
Name string `json:"name"`
}
// AgentNetworkModelDiscoveryRequest defines model for AgentNetworkModelDiscoveryRequest.
type AgentNetworkModelDiscoveryRequest struct {
// ApiKey Credential to query the vendor with, for a provider that has not been saved yet. Mutually exclusive with provider_id.
ApiKey *string `json:"api_key,omitempty"`
// CatalogProviderId Catalog provider to query (AgentNetworkCatalogProvider.id). Determines the listing endpoint, the auth header and the response shape.
CatalogProviderId string `json:"catalog_provider_id"`
// ProviderId Existing Agent Network provider record whose stored credential and upstream should be used. Lets the form refresh the list without the client holding the key.
ProviderId *string `json:"provider_id,omitempty"`
// UpstreamUrl The upstream being configured. Used to reach vendors that serve their listing from the same host as inference, and to read back the region for those whose host embeds one. Ignored when provider_id is supplied.
UpstreamUrl *string `json:"upstream_url,omitempty"`
}
// AgentNetworkModelDiscoveryResponse defines model for AgentNetworkModelDiscoveryResponse.
type AgentNetworkModelDiscoveryResponse struct {
// Models Models the credential can reach, in the order the vendor returned them.
Models []AgentNetworkDiscoveredModel `json:"models"`
}
// AgentNetworkPolicy defines model for AgentNetworkPolicy.
type AgentNetworkPolicy struct {
// CreatedAt Timestamp when the policy was created.
@@ -2876,7 +2924,7 @@ type EDRFleetDMRequest struct {
// LastSyncedInterval The devices last sync requirement interval in hours. Minimum value is 24 hours
LastSyncedInterval int `json:"last_synced_interval"`
// MatchAttributes Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open-source version. Premium-only attributes are marked accordingly
// MatchAttributes Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open source version. Premium-only attributes are marked accordingly
MatchAttributes FleetDMMatchAttributes `json:"match_attributes"`
}
@@ -2909,7 +2957,7 @@ type EDRFleetDMResponse struct {
// LastSyncedInterval The devices last sync requirement interval in hours.
LastSyncedInterval int `json:"last_synced_interval"`
// MatchAttributes Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open-source version. Premium-only attributes are marked accordingly
// MatchAttributes Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open source version. Premium-only attributes are marked accordingly
MatchAttributes FleetDMMatchAttributes `json:"match_attributes"`
// UpdatedAt Timestamp of when the integration was last updated.
@@ -3129,7 +3177,7 @@ type Event struct {
// EventActivityCode The string code of the activity that occurred during the event
type EventActivityCode string
// FleetDMMatchAttributes Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open-source version. Premium-only attributes are marked accordingly
// FleetDMMatchAttributes Attribute conditions to match when approving FleetDM hosts. Most attributes work with FleetDM's free/open source version. Premium-only attributes are marked accordingly
type FleetDMMatchAttributes struct {
// DiskEncryptionEnabled Whether disk encryption (FileVault/BitLocker) must be enabled on the host
DiskEncryptionEnabled *bool `json:"disk_encryption_enabled,omitempty"`
@@ -6179,6 +6227,9 @@ type PostApiAgentNetworkBudgetRulesJSONRequestBody = AgentNetworkBudgetRuleReque
// PutApiAgentNetworkBudgetRulesRuleIdJSONRequestBody defines body for PutApiAgentNetworkBudgetRulesRuleId for application/json ContentType.
type PutApiAgentNetworkBudgetRulesRuleIdJSONRequestBody = AgentNetworkBudgetRuleRequest
// PostApiAgentNetworkCatalogProvidersModelsJSONRequestBody defines body for PostApiAgentNetworkCatalogProvidersModels for application/json ContentType.
type PostApiAgentNetworkCatalogProvidersModelsJSONRequestBody = AgentNetworkModelDiscoveryRequest
// PostApiAgentNetworkGuardrailsJSONRequestBody defines body for PostApiAgentNetworkGuardrails for application/json ContentType.
type PostApiAgentNetworkGuardrailsJSONRequestBody = AgentNetworkGuardrailRequest
+1
View File
@@ -275,6 +275,7 @@ func decodePeerCompact(pc *proto.PeerCompact, peerID string) *types.ComponentPee
SupportsIPv6: pc.SupportsIpv6,
ServerSSHAllowed: pc.ServerSshAllowed,
AddedWithSSOLogin: pc.AddedWithSsoLogin,
ProxyEmbedded: pc.ProxyEmbedded,
}
if pc.LastLoginUnixNano != 0 {
peer.LastLogin = time.Unix(0, pc.LastLoginUnixNano)
+17 -3
View File
@@ -247,7 +247,7 @@ func ToProtocolDNSConfig(update nbdns.Config, cache DNSConfigCache, forwardPort
ServiceEnable: update.ServiceEnable,
CustomZones: make([]*proto.CustomZone, 0, len(update.CustomZones)),
NameServerGroups: make([]*proto.NameServerGroup, 0, len(update.NameServerGroups)),
ForwarderPort: forwardPort,
ForwarderPort: forwardPort, //nolint:staticcheck
}
for _, zone := range update.CustomZones {
@@ -272,8 +272,9 @@ func ToProtocolDNSConfig(update nbdns.Config, cache DNSConfigCache, forwardPort
}
// AppendRemotePeerConfig appends typed peers as proto.RemotePeerConfig
// entries to dst and returns the result.
func AppendRemotePeerConfig(dst []*proto.RemotePeerConfig, peers []*types.ComponentPeer, dnsName string, includeIPv6 bool) []*proto.RemotePeerConfig {
// entries to dst and returns the result. localIsProxy reports whether the peer
// receiving this config is itself an embedded proxy.
func AppendRemotePeerConfig(dst []*proto.RemotePeerConfig, peers []*types.ComponentPeer, dnsName string, includeIPv6 bool, localIsProxy bool) []*proto.RemotePeerConfig {
for _, rPeer := range peers {
allowedIPs := []string{rPeer.IP.String() + "/32"}
if includeIPv6 && rPeer.IPv6.IsValid() {
@@ -285,11 +286,24 @@ func AppendRemotePeerConfig(dst []*proto.RemotePeerConfig, peers []*types.Compon
SshConfig: &proto.SSHConfig{SshPubKey: []byte(rPeer.SSHKey)},
Fqdn: rPeer.FQDN(dnsName),
AgentVersion: rPeer.AgentVersion,
LazyState: lazyStateFor(localIsProxy, rPeer),
})
}
return dst
}
// lazyStateFor returns the per-peer lazy override for a remote peer. Connections
// involving an ephemeral proxy peer on either endpoint default to lazy so shared
// proxy infrastructure is not kept permanently connected to every peer. All
// other peers follow the account-wide flag. A future admin-facing per-peer
// setting can return LazyStateEager here to force a peer always-active.
func lazyStateFor(localIsProxy bool, rPeer *types.ComponentPeer) proto.LazyState {
if localIsProxy || rPeer.ProxyEmbedded {
return proto.LazyState_LazyStateLazy
}
return proto.LazyState_LazyStateDefault
}
// BuildAuthorizedUsersProto deduplicates user-IDs into a hashed list and
// builds per-machine-user index maps. Returns (hashedUsers, machineUsers).
// Errors from individual hash failures are logged via the provided context;
+2 -2
View File
@@ -74,11 +74,11 @@ func EnvelopeToNetworkMap(ctx context.Context, env *proto.NetworkMapEnvelope, lo
protoNM.Routes = ToProtocolRoutes(typedNM.Routes)
protoNM.DNSConfig = ToProtocolDNSConfig(typedNM.DNSConfig, nil, dnsFwdPort)
remotePeers := AppendRemotePeerConfig(nil, typedNM.Peers, dnsName, includeIPv6)
remotePeers := AppendRemotePeerConfig(nil, typedNM.Peers, dnsName, includeIPv6, localPeer.ProxyEmbedded)
protoNM.RemotePeers = remotePeers
protoNM.RemotePeersIsEmpty = len(remotePeers) == 0
protoNM.OfflinePeers = AppendRemotePeerConfig(nil, typedNM.OfflinePeers, dnsName, includeIPv6)
protoNM.OfflinePeers = AppendRemotePeerConfig(nil, typedNM.OfflinePeers, dnsName, includeIPv6, localPeer.ProxyEmbedded)
firewallRules := ToProtocolFirewallRules(typedNM.FirewallRules, includeIPv6, useSourcePrefixes)
protoNM.FirewallRules = firewallRules
File diff suppressed because it is too large Load Diff
+21
View File
@@ -501,6 +501,22 @@ message RemotePeerConfig {
string fqdn = 4;
string agentVersion = 5;
// lazyState is the management per-peer override for lazy (on-demand)
// connections to this remote peer. LazyStateDefault follows the account-wide
// flag; LazyStateLazy forces lazy; LazyStateEager forces an always-active
// connection. A local NB_LAZY_CONN/MDM override still wins over this.
LazyState lazyState = 6;
}
// LazyState is the management per-peer override for lazy connections.
enum LazyState {
// Follow the account-wide lazy connection flag.
LazyStateDefault = 0;
// Force a lazy (on-demand) connection regardless of the account flag.
LazyStateLazy = 1;
// Force an always-active connection regardless of the account flag.
LazyStateEager = 2;
}
// SSHConfig represents SSH configurations of a peer.
@@ -1026,6 +1042,11 @@ message PeerCompact {
// (port 22022) is only added when this flag is set and the peer agent
// version supports it.
bool server_ssh_allowed = 13;
// Mirror of types.Peer.ProxyMeta.Embedded. Connections involving an
// ephemeral proxy peer on either endpoint default to lazy, so this bit
// feeds the per-peer lazyState emitted in RemotePeerConfig.
bool proxy_embedded = 14;
}
// PolicyCompact is the compact form of a policy rule. Group references use
@@ -25,6 +25,9 @@ type ComponentPeer struct {
LoginExpirationEnabled bool
AddedWithSSOLogin bool
LastLogin time.Time
// ProxyEmbedded marks an ephemeral embedded proxy peer. Connections
// involving such a peer on either endpoint default to lazy.
ProxyEmbedded bool
}
// FQDN returns the peer's FQDN combined of the peer's DNS label and the system's DNS domain.