[management, client, proxy] add expose NetBird-only services over tunnel peers (#6226)

Adds a new "private" service mode for the reverse proxy: services reachable exclusively over the embedded WireGuard tunnel, gated by per-peer group membership instead of operator auth schemes.

Wire contract
- ProxyMapping.private (field 13): the proxy MUST call ValidateTunnelPeer and fail closed; operator schemes are bypassed.
- ProxyCapabilities.private (4) + supports_private_service (5): capability gate. Management never streams private mappings to proxies that don't claim the capability; the broadcast path applies the same filter via filterMappingsForProxy.
- ValidateTunnelPeer RPC: resolves an inbound tunnel IP to a peer, checks the peer's groups against service.AccessGroups, and mints a session JWT on success. checkPeerGroupAccess fails closed when a private service has empty AccessGroups.
- ValidateSession/ValidateTunnelPeer responses now carry peer_group_ids + peer_group_names so the proxy can authorise policy-aware middlewares without an extra management round-trip.
- ProxyInboundListener + SendStatusUpdate.inbound_listener: per-account inbound listener state surfaced to dashboards.
- PathTargetOptions.direct_upstream (11): bypass the embedded NetBird client and dial the target via the proxy host's network stack for upstreams reachable without WireGuard.

Data model
- Service.Private (bool) + Service.AccessGroups ([]string, JSON- serialised). Validate() rejects bearer auth on private services. Copy() deep-copies AccessGroups. pgx getServices loads the columns.
- DomainConfig.Private threaded into the proxy auth middleware. Request handler routes private services through forwardWithTunnelPeer and returns 403 on validation failure.
- Account-level SynthesizePrivateServiceZones (synthetic DNS) and injectPrivateServicePolicies (synthetic ACL) gate on len(svc.AccessGroups) > 0.

Proxy
- /netbird proxy --private (embedded mode) flag; Config.Private in proxy/lifecycle.go.
- Per-account inbound listener (proxy/inbound.go) binding HTTP/HTTPS on the embedded NetBird client's WireGuard tunnel netstack.
- proxy/internal/auth/tunnel_cache: ValidateTunnelPeer response cache with single-flight de-duplication and per-account eviction.
- Local peerstore short-circuit: when the inbound IP isn't in the account roster, deny fast without an RPC.
- proxy/server.go reports SupportsPrivateService=true and redacts the full ProxyMapping JSON from info logs (auth_token + header-auth hashed values now only at debug level).

Identity forwarding
- ValidateSessionJWT returns user_id, email, method, groups, group_names. sessionkey.Claims carries Email + Groups + GroupNames so the proxy can stamp identity onto upstream requests without an extra management round-trip on every cookie-bearing request.
- CapturedData carries userEmail / userGroups / userGroupNames; the proxy stamps X-NetBird-User and X-NetBird-Groups on r.Out from the authenticated identity (strips client-supplied values first to prevent spoofing).
- AccessLog.UserGroups: access-log enrichment captures the user's group memberships at write time so the dashboard can render group context without reverse-resolving stale memberships.

OpenAPI/dashboard surface
- ReverseProxyService gains private + access_groups; ReverseProxyCluster gains private + supports_private. ReverseProxyTarget target_type enum gains "cluster". ServiceTargetOptions gains direct_upstream. ProxyAccessLog gains user_groups.
This commit is contained in:
Maycon Santos
2026-05-25 17:41:50 +02:00
committed by GitHub
parent 0358be2313
commit 7aebdd69dd
84 changed files with 7810 additions and 933 deletions
@@ -23,6 +23,8 @@ type Domain struct {
// SupportsCrowdSec is populated at query time from proxy cluster capabilities.
// Not persisted.
SupportsCrowdSec *bool `gorm:"-"`
// SupportsPrivate is populated at query time from proxy cluster capabilities. Not persisted.
SupportsPrivate *bool `gorm:"-"`
}
// EventMeta returns activity event metadata for a domain
@@ -49,6 +49,7 @@ func domainToApi(d *domain.Domain) api.ReverseProxyDomain {
SupportsCustomPorts: d.SupportsCustomPorts,
RequireSubdomain: d.RequireSubdomain,
SupportsCrowdsec: d.SupportsCrowdSec,
SupportsPrivate: d.SupportsPrivate,
}
if d.TargetCluster != "" {
resp.TargetCluster = &d.TargetCluster
@@ -35,6 +35,7 @@ type proxyManager interface {
ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
}
type Manager struct {
@@ -93,6 +94,7 @@ func (m Manager) GetDomains(ctx context.Context, accountID, userID string) ([]*d
d.SupportsCustomPorts = m.proxyManager.ClusterSupportsCustomPorts(ctx, cluster)
d.RequireSubdomain = m.proxyManager.ClusterRequireSubdomain(ctx, cluster)
d.SupportsCrowdSec = m.proxyManager.ClusterSupportsCrowdSec(ctx, cluster)
d.SupportsPrivate = m.proxyManager.ClusterSupportsPrivate(ctx, cluster)
ret = append(ret, d)
}
@@ -109,6 +111,7 @@ func (m Manager) GetDomains(ctx context.Context, accountID, userID string) ([]*d
if d.TargetCluster != "" {
cd.SupportsCustomPorts = m.proxyManager.ClusterSupportsCustomPorts(ctx, d.TargetCluster)
cd.SupportsCrowdSec = m.proxyManager.ClusterSupportsCrowdSec(ctx, d.TargetCluster)
cd.SupportsPrivate = m.proxyManager.ClusterSupportsPrivate(ctx, d.TargetCluster)
}
// Custom domains never require a subdomain by default since
// the account owns them and should be able to use the bare domain.
@@ -10,7 +10,7 @@ import (
)
type mockProxyManager struct {
getActiveClusterAddressesFunc func(ctx context.Context) ([]string, error)
getActiveClusterAddressesFunc func(ctx context.Context) ([]string, error)
getActiveClusterAddressesForAccountFunc func(ctx context.Context, accountID string) ([]string, error)
}
@@ -40,6 +40,10 @@ func (m *mockProxyManager) ClusterSupportsCrowdSec(_ context.Context, _ string)
return nil
}
func (m *mockProxyManager) ClusterSupportsPrivate(_ context.Context, _ string) *bool {
return nil
}
func TestGetClusterAllowList_BYOPMergedWithPublic(t *testing.T) {
pm := &mockProxyManager{
getActiveClusterAddressesForAccountFunc: func(_ context.Context, accID string) ([]string, error) {
@@ -151,4 +155,3 @@ func TestGetClusterAllowList_PublicEmpty_BYOPOnly(t *testing.T) {
require.NoError(t, err)
assert.Equal(t, []string{"byop.example.com"}, result)
}
@@ -19,6 +19,7 @@ type Manager interface {
ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
CleanupStale(ctx context.Context, inactivityDuration time.Duration) error
GetAccountProxy(ctx context.Context, accountID string) (*Proxy, error)
CountAccountProxies(ctx context.Context, accountID string) (int64, error)
@@ -21,6 +21,7 @@ type store interface {
GetClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
GetClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
GetClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
GetClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
CleanupStaleProxies(ctx context.Context, inactivityDuration time.Duration) error
GetProxyByAccountID(ctx context.Context, accountID string) (*proxy.Proxy, error)
CountProxiesByAccountID(ctx context.Context, accountID string) (int64, error)
@@ -137,6 +138,11 @@ func (m Manager) ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string
return m.store.GetClusterSupportsCrowdSec(ctx, clusterAddr)
}
// ClusterSupportsPrivate reports whether any active proxy claims the private capability (nil = unreported).
func (m Manager) ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool {
return m.store.GetClusterSupportsPrivate(ctx, clusterAddr)
}
// CleanupStale removes proxies that haven't sent heartbeat in the specified duration
func (m *Manager) CleanupStale(ctx context.Context, inactivityDuration time.Duration) error {
if err := m.store.CleanupStaleProxies(ctx, inactivityDuration); err != nil {
@@ -178,4 +184,3 @@ func (m *Manager) DeleteAccountCluster(ctx context.Context, clusterAddress, acco
}
return nil
}
@@ -15,16 +15,16 @@ import (
)
type mockStore struct {
saveProxyFunc func(ctx context.Context, p *proxy.Proxy) error
disconnectProxyFunc func(ctx context.Context, proxyID, sessionID string) error
updateProxyHeartbeatFunc func(ctx context.Context, p *proxy.Proxy) error
getActiveProxyClusterAddressesFunc func(ctx context.Context) ([]string, error)
getActiveProxyClusterAddressesForAccFunc func(ctx context.Context, accountID string) ([]string, error)
cleanupStaleProxiesFunc func(ctx context.Context, d time.Duration) error
getProxyByAccountIDFunc func(ctx context.Context, accountID string) (*proxy.Proxy, error)
countProxiesByAccountIDFunc func(ctx context.Context, accountID string) (int64, error)
isClusterAddressConflictingFunc func(ctx context.Context, clusterAddress, accountID string) (bool, error)
deleteAccountClusterFunc func(ctx context.Context, clusterAddress, accountID string) error
saveProxyFunc func(ctx context.Context, p *proxy.Proxy) error
disconnectProxyFunc func(ctx context.Context, proxyID, sessionID string) error
updateProxyHeartbeatFunc func(ctx context.Context, p *proxy.Proxy) error
getActiveProxyClusterAddressesFunc func(ctx context.Context) ([]string, error)
getActiveProxyClusterAddressesForAccFunc func(ctx context.Context, accountID string) ([]string, error)
cleanupStaleProxiesFunc func(ctx context.Context, d time.Duration) error
getProxyByAccountIDFunc func(ctx context.Context, accountID string) (*proxy.Proxy, error)
countProxiesByAccountIDFunc func(ctx context.Context, accountID string) (int64, error)
isClusterAddressConflictingFunc func(ctx context.Context, clusterAddress, accountID string) (bool, error)
deleteAccountClusterFunc func(ctx context.Context, clusterAddress, accountID string) error
}
func (m *mockStore) SaveProxy(ctx context.Context, p *proxy.Proxy) error {
@@ -99,6 +99,9 @@ func (m *mockStore) GetClusterRequireSubdomain(_ context.Context, _ string) *boo
func (m *mockStore) GetClusterSupportsCrowdSec(_ context.Context, _ string) *bool {
return nil
}
func (m *mockStore) GetClusterSupportsPrivate(_ context.Context, _ string) *bool {
return nil
}
func newTestManager(s store) *Manager {
meter := noop.NewMeterProvider().Meter("test")
@@ -92,6 +92,20 @@ func (mr *MockManagerMockRecorder) ClusterSupportsCrowdSec(ctx, clusterAddr inte
return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCrowdSec", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCrowdSec), ctx, clusterAddr)
}
// ClusterSupportsPrivate mocks base method.
func (m *MockManager) ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool {
m.ctrl.T.Helper()
ret := m.ctrl.Call(m, "ClusterSupportsPrivate", ctx, clusterAddr)
ret0, _ := ret[0].(*bool)
return ret0
}
// ClusterSupportsPrivate indicates an expected call of ClusterSupportsPrivate.
func (mr *MockManagerMockRecorder) ClusterSupportsPrivate(ctx, clusterAddr interface{}) *gomock.Call {
mr.mock.ctrl.T.Helper()
return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsPrivate", reflect.TypeOf((*MockManager)(nil).ClusterSupportsPrivate), ctx, clusterAddr)
}
// Connect mocks base method.
func (m *MockManager) Connect(ctx context.Context, proxyID, sessionID, clusterAddress, ipAddress string, accountID *string, capabilities *Capabilities) (*Proxy, error) {
m.ctrl.T.Helper()
@@ -20,6 +20,9 @@ type Capabilities struct {
RequireSubdomain *bool
// SupportsCrowdsec indicates whether this proxy has CrowdSec configured.
SupportsCrowdsec *bool
// Private indicates whether this proxy supports inbound access via Wireguard
// tunnel and netbird-only authentication policies
Private *bool
}
// Proxy represents a reverse proxy instance
@@ -67,10 +70,9 @@ type Cluster struct {
Type ClusterType
Online bool
ConnectedProxies int
// Capability flags. *bool because nil means "no proxy reported a
// capability for this cluster" — the dashboard renders these as
// unknown rather than false.
// *bool: nil = no proxy reported the capability; the dashboard renders that as unknown.
SupportsCustomPorts *bool
RequireSubdomain *bool
SupportsCrowdSec *bool
Private *bool
}
@@ -204,6 +204,7 @@ func (h *handler) getClusters(w http.ResponseWriter, r *http.Request) {
SupportsCustomPorts: c.SupportsCustomPorts,
RequireSubdomain: c.RequireSubdomain,
SupportsCrowdsec: c.SupportsCrowdSec,
Private: c.Private,
})
}
@@ -82,6 +82,7 @@ type CapabilityProvider interface {
ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
}
type Manager struct {
@@ -136,6 +137,7 @@ func (m *Manager) GetClusters(ctx context.Context, accountID, userID string) ([]
clusters[i].SupportsCustomPorts = m.capabilities.ClusterSupportsCustomPorts(ctx, clusters[i].Address)
clusters[i].RequireSubdomain = m.capabilities.ClusterRequireSubdomain(ctx, clusters[i].Address)
clusters[i].SupportsCrowdSec = m.capabilities.ClusterSupportsCrowdSec(ctx, clusters[i].Address)
clusters[i].Private = m.capabilities.ClusterSupportsPrivate(ctx, clusters[i].Address)
}
return clusters, nil
@@ -208,6 +210,9 @@ func (m *Manager) replaceHostByLookup(ctx context.Context, accountID string, s *
target.Host = resource.Domain
case service.TargetTypeSubnet:
// For subnets we do not do any lookups on the resource
case service.TargetTypeCluster:
// Cluster targets carry the upstream address on target_id; the
// proxy resolves the destination at request time.
default:
return fmt.Errorf("unknown target type: %s", target.TargetType)
}
@@ -779,6 +784,10 @@ func validateTargetReferences(ctx context.Context, transaction store.Store, acco
if err := validateResourceTarget(ctx, transaction, accountID, target); err != nil {
return err
}
case service.TargetTypeCluster:
if err := validateClusterTarget(target); err != nil {
return err
}
default:
return status.Errorf(status.InvalidArgument, "unknown target type %q for target %q", target.TargetType, target.TargetId)
}
@@ -786,6 +795,13 @@ func validateTargetReferences(ctx context.Context, transaction store.Store, acco
return nil
}
func validateClusterTarget(target *service.Target) error {
if !target.Options.DirectUpstream {
return status.Errorf(status.InvalidArgument, "cluster target %s has direct upstream disabled", target.Host)
}
return nil
}
func validatePeerTarget(ctx context.Context, transaction store.Store, accountID string, target *service.Target) error {
if _, err := transaction.GetPeerByID(ctx, store.LockingStrengthShare, accountID, target.TargetId); err != nil {
if sErr, ok := status.FromError(err); ok && sErr.Type() == status.NotFound {
@@ -962,12 +978,14 @@ func (m *Manager) ReloadAllServicesForAccount(ctx context.Context, accountID str
return fmt.Errorf("failed to get services: %w", err)
}
oidcCfg := m.proxyController.GetOIDCValidationConfig()
for _, s := range services {
err = m.replaceHostByLookup(ctx, accountID, s)
if err != nil {
return fmt.Errorf("failed to replace host by lookup for service %s: %w", s.ID, err)
}
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, s.ToProtoMapping(service.Update, "", m.proxyController.GetOIDCValidationConfig()), s.ProxyCluster)
m.proxyController.SendServiceUpdateToCluster(ctx, accountID, s.ToProtoMapping(service.Update, "", oidcCfg), s.ProxyCluster)
}
return nil
@@ -1344,3 +1344,66 @@ func TestValidateSubdomainRequirement(t *testing.T) {
})
}
}
func TestValidateTargetReferences_ClusterTargetSkipsLookup(t *testing.T) {
ctx := context.Background()
ctrl := gomock.NewController(t)
mockStore := store.NewMockStore(ctrl)
accountID := "test-account"
// No peer or resource lookups must be issued for cluster targets.
targets := []*rpservice.Target{
{
TargetId: "eu.proxy.netbird.io",
TargetType: rpservice.TargetTypeCluster,
Options: rpservice.TargetOptions{DirectUpstream: true},
},
}
require.NoError(t, validateTargetReferences(ctx, mockStore, accountID, targets), "cluster target must validate without store lookups")
}
// TestValidateTargetReferences_ClusterTargetRequiresDirectUpstream pins the
// store-side check that cluster targets must opt into the host-stack dial
// path. Without DirectUpstream the proxy would route this target through
// the embedded NetBird client and fail on every request.
func TestValidateTargetReferences_ClusterTargetRequiresDirectUpstream(t *testing.T) {
ctx := context.Background()
ctrl := gomock.NewController(t)
mockStore := store.NewMockStore(ctrl)
accountID := "test-account"
targets := []*rpservice.Target{
{
TargetId: "eu.proxy.netbird.io",
TargetType: rpservice.TargetTypeCluster,
Host: "backend.lan",
},
}
err := validateTargetReferences(ctx, mockStore, accountID, targets)
require.Error(t, err, "cluster target without direct_upstream must be rejected")
assert.ErrorContains(t, err, "direct upstream disabled")
}
func TestReplaceHostByLookup_SkipsClusterTarget(t *testing.T) {
ctx := context.Background()
ctrl := gomock.NewController(t)
mockStore := store.NewMockStore(ctrl)
accountID := "test-account"
mgr := &Manager{store: mockStore}
svc := &rpservice.Service{
ID: "svc-1",
AccountID: accountID,
Targets: []*rpservice.Target{
{
TargetId: "eu.proxy.netbird.io",
TargetType: rpservice.TargetTypeCluster,
Host: "127.0.0.1",
},
},
}
require.NoError(t, mgr.replaceHostByLookup(ctx, accountID, svc), "cluster target must not trigger peer/resource lookup")
assert.Equal(t, "127.0.0.1", svc.Targets[0].Host, "operator-supplied host must be preserved for cluster target")
}
@@ -45,10 +45,11 @@ const (
StatusCertificateFailed Status = "certificate_failed"
StatusError Status = "error"
TargetTypePeer TargetType = "peer"
TargetTypeHost TargetType = "host"
TargetTypeDomain TargetType = "domain"
TargetTypeSubnet TargetType = "subnet"
TargetTypePeer TargetType = "peer"
TargetTypeHost TargetType = "host"
TargetTypeDomain TargetType = "domain"
TargetTypeSubnet TargetType = "subnet"
TargetTypeCluster TargetType = "cluster"
SourcePermanent = "permanent"
SourceEphemeral = "ephemeral"
@@ -60,6 +61,11 @@ type TargetOptions struct {
SessionIdleTimeout time.Duration `json:"session_idle_timeout,omitempty"`
PathRewrite PathRewriteMode `json:"path_rewrite,omitempty"`
CustomHeaders map[string]string `gorm:"serializer:json" json:"custom_headers,omitempty"`
// DirectUpstream bypasses the proxy's embedded NetBird client and dials
// the target via the proxy host's network stack. Useful for upstreams
// reachable without WireGuard (public APIs, LAN services, localhost
// sidecars). Default false.
DirectUpstream bool `json:"direct_upstream,omitempty"`
}
type Target struct {
@@ -67,7 +73,7 @@ type Target struct {
AccountID string `gorm:"index:idx_target_account;not null" json:"-"`
ServiceID string `gorm:"index:idx_service_targets;not null" json:"-"`
Path *string `json:"path,omitempty"`
Host string `json:"host"` // the Host field is only used for subnet targets, otherwise ignored
Host string `json:"host"`
Port uint16 `gorm:"index:idx_target_port" json:"port"`
Protocol string `gorm:"index:idx_target_protocol" json:"protocol"`
TargetId string `gorm:"index:idx_target_id" json:"target_id"`
@@ -200,6 +206,10 @@ type Service struct {
Mode string `gorm:"default:'http'"`
ListenPort uint16
PortAutoAssigned bool
// Private marks the service as NetBird-only: auth via ValidateTunnelPeer against AccessGroups instead of SSO. HTTP-only.
Private bool
// AccessGroups is the group ID allowlist for inbound peers on private services. Mutually exclusive with bearer SSO.
AccessGroups []string `json:"access_groups,omitempty" gorm:"serializer:json"`
}
// InitNewRecord generates a new unique ID and resets metadata for a newly created
@@ -299,6 +309,12 @@ func (s *Service) ToAPIResponse() *api.Service {
Mode: &mode,
ListenPort: &listenPort,
PortAutoAssigned: &s.PortAutoAssigned,
Private: &s.Private,
}
if len(s.AccessGroups) > 0 {
groups := append([]string(nil), s.AccessGroups...)
resp.AccessGroups = &groups
}
if s.ProxyCluster != "" {
@@ -308,6 +324,7 @@ func (s *Service) ToAPIResponse() *api.Service {
return resp
}
// ToProtoMapping converts the service into the wire format the proxy consumes.
func (s *Service) ToProtoMapping(operation Operation, authToken string, oidcConfig proxy.OIDCValidationConfig) *proto.ProxyMapping {
pathMappings := s.buildPathMappings()
@@ -349,6 +366,7 @@ func (s *Service) ToProtoMapping(operation Operation, authToken string, oidcConf
RewriteRedirects: s.RewriteRedirects,
Mode: s.Mode,
ListenPort: int32(s.ListenPort), //nolint:gosec
Private: s.Private,
}
if r := restrictionsToProto(s.Restrictions); r != nil {
@@ -455,7 +473,8 @@ func pathRewriteToProto(mode PathRewriteMode) proto.PathRewriteMode {
}
func targetOptionsToAPI(opts TargetOptions) *api.ServiceTargetOptions {
if !opts.SkipTLSVerify && opts.RequestTimeout == 0 && opts.SessionIdleTimeout == 0 && opts.PathRewrite == "" && len(opts.CustomHeaders) == 0 {
if !opts.SkipTLSVerify && opts.RequestTimeout == 0 && opts.SessionIdleTimeout == 0 &&
opts.PathRewrite == "" && len(opts.CustomHeaders) == 0 && !opts.DirectUpstream {
return nil
}
apiOpts := &api.ServiceTargetOptions{}
@@ -477,17 +496,22 @@ func targetOptionsToAPI(opts TargetOptions) *api.ServiceTargetOptions {
if len(opts.CustomHeaders) > 0 {
apiOpts.CustomHeaders = &opts.CustomHeaders
}
if opts.DirectUpstream {
apiOpts.DirectUpstream = &opts.DirectUpstream
}
return apiOpts
}
func targetOptionsToProto(opts TargetOptions) *proto.PathTargetOptions {
if !opts.SkipTLSVerify && opts.PathRewrite == "" && opts.RequestTimeout == 0 && len(opts.CustomHeaders) == 0 {
if !opts.SkipTLSVerify && opts.PathRewrite == "" && opts.RequestTimeout == 0 &&
len(opts.CustomHeaders) == 0 && !opts.DirectUpstream {
return nil
}
popts := &proto.PathTargetOptions{
SkipTlsVerify: opts.SkipTLSVerify,
PathRewrite: pathRewriteToProto(opts.PathRewrite),
CustomHeaders: opts.CustomHeaders,
SkipTlsVerify: opts.SkipTLSVerify,
PathRewrite: pathRewriteToProto(opts.PathRewrite),
CustomHeaders: opts.CustomHeaders,
DirectUpstream: opts.DirectUpstream,
}
if opts.RequestTimeout != 0 {
popts.RequestTimeout = durationpb.New(opts.RequestTimeout)
@@ -537,6 +561,9 @@ func targetOptionsFromAPI(idx int, o *api.ServiceTargetOptions) (TargetOptions,
if o.CustomHeaders != nil {
opts.CustomHeaders = *o.CustomHeaders
}
if o.DirectUpstream != nil {
opts.DirectUpstream = *o.DirectUpstream
}
return opts, nil
}
@@ -551,6 +578,14 @@ func (s *Service) FromAPIRequest(req *api.ServiceRequest, accountID string) erro
if req.ListenPort != nil {
s.ListenPort = uint16(*req.ListenPort) //nolint:gosec
}
if req.Private != nil {
s.Private = *req.Private
}
if req.AccessGroups != nil {
s.AccessGroups = append([]string(nil), *req.AccessGroups...)
} else {
s.AccessGroups = nil
}
targets, err := targetsFromAPI(accountID, req.Targets)
if err != nil {
@@ -740,6 +775,9 @@ func (s *Service) Validate() error {
if err := validateAccessRestrictions(&s.Restrictions); err != nil {
return err
}
if err := s.validatePrivateRequirements(); err != nil {
return err
}
switch s.Mode {
case ModeHTTP:
@@ -753,6 +791,23 @@ func (s *Service) Validate() error {
}
}
// validatePrivateRequirements enforces the private-service contract: HTTP mode, ≥1 access group, no bearer auth.
func (s *Service) validatePrivateRequirements() error {
if !s.Private {
return nil
}
if s.Mode != "" && s.Mode != ModeHTTP {
return fmt.Errorf("private services only support HTTP mode, got %q", s.Mode)
}
if len(s.AccessGroups) == 0 {
return errors.New("private services require at least one access group")
}
if s.Auth.BearerAuth != nil && s.Auth.BearerAuth.Enabled {
return errors.New("private services cannot enable bearer auth (SSO): NetBird-only access and SSO are mutually exclusive")
}
return nil
}
func (s *Service) validateHTTPMode() error {
if s.Domain == "" {
return errors.New("service domain is required")
@@ -799,11 +854,21 @@ func (s *Service) validateHTTPTargets() error {
for i, target := range s.Targets {
switch target.TargetType {
case TargetTypePeer, TargetTypeHost, TargetTypeDomain:
// host field will be ignored
// Host is normally overwritten by replaceHostByLookup with the
// resolved peer IP / resource address; operator-supplied values
// are honored only when DirectUpstream is set. Validate the
// override here so misconfigured hosts fail fast at API time.
if err := validateDirectUpstreamHost(i, target); err != nil {
return err
}
case TargetTypeSubnet:
if target.Host == "" {
return fmt.Errorf("target %d has empty host but target_type is %q", i, target.TargetType)
}
case TargetTypeCluster:
if err := validateClusterTarget(i, target); err != nil {
return err
}
default:
return fmt.Errorf("target %d has invalid target_type %q", i, target.TargetType)
}
@@ -821,25 +886,67 @@ func (s *Service) validateHTTPTargets() error {
return nil
}
// validateClusterTarget cluster targets should not have empty hosts and should have direct upstream enabled.
func validateClusterTarget(idx int, target *Target) error {
host := strings.TrimSpace(target.Host)
if host == "" {
return fmt.Errorf("target %d: has empty host", idx)
}
if !target.Options.DirectUpstream {
return fmt.Errorf("target %d: %s has direct upstream disabled", idx, target.Host)
}
return validateDirectUpstreamHost(idx, target)
}
// validateDirectUpstreamHost validates the operator-supplied Host on a
// peer/host/domain target when DirectUpstream is set. Empty Host is
// allowed — the lookup fills in the default peer IP / resource address.
// Without DirectUpstream the Host value is silently overwritten by
// replaceHostByLookup, so we don't validate it (preserves the historical
// behaviour where APIs accepted any value and dropped it). Non-empty
// Host with DirectUpstream must look like a hostname or IP and must
// not carry a port (port lives on Target.Port).
func validateDirectUpstreamHost(idx int, target *Target) error {
if !target.Options.DirectUpstream {
return nil
}
host := strings.TrimSpace(target.Host)
if host == "" {
return nil
}
if strings.ContainsAny(host, " \t/") {
return fmt.Errorf("target %d: host %q contains invalid characters", idx, host)
}
if _, _, err := net.SplitHostPort(host); err == nil {
return fmt.Errorf("target %d: host %q must not include a port (set target.port instead)", idx, host)
}
return nil
}
func (s *Service) validateL4Target(target *Target) error {
// L4 services have a single target; per-target disable is meaningless
// (use the service-level Enabled flag instead). Force it on so that
// buildPathMappings always includes the target in the proto.
target.Enabled = true
if target.Port == 0 {
return errors.New("target port is required for L4 services")
}
if target.TargetId == "" {
return errors.New("target_id is required for L4 services")
}
if target.TargetType != TargetTypeCluster && target.Port == 0 {
return errors.New("target port is required for L4 services")
}
switch target.TargetType {
case TargetTypePeer, TargetTypeHost, TargetTypeDomain:
// OK
if err := validateDirectUpstreamHost(0, target); err != nil {
return err
}
case TargetTypeSubnet:
if target.Host == "" {
return errors.New("target host is required for subnet targets")
}
case TargetTypeCluster:
// target_id carries the cluster address; the proxy resolves
// the upstream at request time.
default:
return fmt.Errorf("invalid target_type %q for L4 service", target.TargetType)
}
@@ -1174,6 +1281,11 @@ func (s *Service) Copy() *Service {
}
}
var accessGroups []string
if len(s.AccessGroups) > 0 {
accessGroups = append([]string(nil), s.AccessGroups...)
}
return &Service{
ID: s.ID,
AccountID: s.AccountID,
@@ -1195,6 +1307,8 @@ func (s *Service) Copy() *Service {
Mode: s.Mode,
ListenPort: s.ListenPort,
PortAutoAssigned: s.PortAutoAssigned,
Private: s.Private,
AccessGroups: accessGroups,
}
}
@@ -12,6 +12,7 @@ import (
"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
"github.com/netbirdio/netbird/shared/hash/argon2id"
"github.com/netbirdio/netbird/shared/management/http/api"
"github.com/netbirdio/netbird/shared/management/proto"
)
@@ -1116,3 +1117,191 @@ func TestValidate_HeaderAuths(t *testing.T) {
assert.Contains(t, err.Error(), "exceeds maximum length")
})
}
func TestValidate_HTTPClusterTarget(t *testing.T) {
rp := validProxy()
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "http",
Host: "backend.lan",
Options: TargetOptions{DirectUpstream: true},
Enabled: true,
}}
require.NoError(t, rp.Validate(), "HTTP cluster target with target_id, host, and direct_upstream must validate")
}
func TestValidate_HTTPClusterTarget_RequiresTargetId(t *testing.T) {
rp := validProxy()
rp.Targets = []*Target{{
TargetType: TargetTypeCluster,
Protocol: "http",
Host: "backend.lan",
Options: TargetOptions{DirectUpstream: true},
Enabled: true,
}}
assert.ErrorContains(t, rp.Validate(), "empty target_id", "cluster target must reject empty target_id")
}
// TestValidate_HTTPClusterTarget_RequiresHost pins the new cluster-target
// rule that operator-supplied Host is mandatory: cluster targets dial the
// upstream via the host network stack (direct_upstream is implied), so an
// empty Host leaves the proxy with nothing to dial.
func TestValidate_HTTPClusterTarget_RequiresHost(t *testing.T) {
rp := validProxy()
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "http",
Options: TargetOptions{DirectUpstream: true},
Enabled: true,
}}
assert.ErrorContains(t, rp.Validate(), "empty host", "cluster target must reject empty host")
}
// TestValidate_HTTPClusterTarget_RequiresDirectUpstream pins the second
// half of the cluster-target rule: DirectUpstream must be true so the
// stdlib transport branch in MultiTransport is taken. Without it the
// embedded NetBird client would try to dial the cluster address through
// the WG tunnel, which is the wrong network for a cluster upstream.
func TestValidate_HTTPClusterTarget_RequiresDirectUpstream(t *testing.T) {
rp := validProxy()
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "http",
Host: "backend.lan",
Enabled: true,
}}
assert.ErrorContains(t, rp.Validate(), "direct upstream disabled", "cluster target must reject direct_upstream=false")
}
func TestValidate_L4ClusterTarget(t *testing.T) {
rp := validProxy()
rp.Mode = ModeTCP
rp.ListenPort = 9000
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "tcp",
Enabled: true,
}}
require.NoError(t, rp.Validate(), "L4 cluster target must validate without an explicit port")
}
func TestService_Copy_RoundtripsPrivate(t *testing.T) {
svc := validProxy()
svc.Private = true
svc.AccessGroups = []string{"grp-admins", "grp-ops"}
cp := svc.Copy()
require.NotNil(t, cp)
assert.True(t, cp.Private)
assert.Equal(t, []string{"grp-admins", "grp-ops"}, cp.AccessGroups)
cp.Private = false
assert.True(t, svc.Private)
cp.AccessGroups[0] = "grp-other"
assert.Equal(t, []string{"grp-admins", "grp-ops"}, svc.AccessGroups)
}
func TestService_APIRoundtrip_Private(t *testing.T) {
enabled := true
private := true
accessGroups := []string{"grp-admins"}
targets := []api.ServiceTarget{{
TargetId: "eu.proxy.netbird.io",
TargetType: api.ServiceTargetTargetType("cluster"),
Protocol: "http",
Port: 80,
Enabled: true,
}}
req := &api.ServiceRequest{
Name: "svc-private",
Domain: "myapp.eu.proxy.netbird.io",
Enabled: enabled,
Private: &private,
AccessGroups: &accessGroups,
Targets: &targets,
}
svc := &Service{}
require.NoError(t, svc.FromAPIRequest(req, "acc-1"))
assert.True(t, svc.Private)
assert.Equal(t, []string{"grp-admins"}, svc.AccessGroups)
resp := svc.ToAPIResponse()
require.NotNil(t, resp.Private)
assert.True(t, *resp.Private)
require.NotNil(t, resp.AccessGroups)
assert.Equal(t, []string{"grp-admins"}, *resp.AccessGroups)
}
func TestValidate_Private_RequiresAccessGroups(t *testing.T) {
rp := validProxy()
rp.Private = true
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "http",
Host: "backend.lan",
Options: TargetOptions{DirectUpstream: true},
Enabled: true,
}}
assert.ErrorContains(t, rp.Validate(), "access group")
}
func TestValidate_Private_RejectsBearerAuth(t *testing.T) {
rp := validProxy()
rp.Private = true
rp.AccessGroups = []string{"grp-admins"}
rp.Auth.BearerAuth = &BearerAuthConfig{
Enabled: true,
DistributionGroups: []string{"grp-sso"},
}
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "http",
Host: "backend.lan",
Options: TargetOptions{DirectUpstream: true},
Enabled: true,
}}
assert.ErrorContains(t, rp.Validate(), "mutually exclusive")
}
func TestValidate_Private_AcceptsNonClusterTargets(t *testing.T) {
rp := validProxy()
rp.Private = true
rp.AccessGroups = []string{"grp-admins"}
require.NoError(t, rp.Validate())
}
func TestValidate_Private_AcceptsClusterTargetWithAccessGroups(t *testing.T) {
rp := validProxy()
rp.Private = true
rp.AccessGroups = []string{"grp-admins"}
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "http",
Host: "backend.lan",
Options: TargetOptions{DirectUpstream: true},
Enabled: true,
}}
require.NoError(t, rp.Validate())
}
func TestValidate_Private_RejectsNonHTTPMode(t *testing.T) {
rp := validProxy()
rp.Private = true
rp.AccessGroups = []string{"grp-admins"}
rp.Mode = ModeTCP
rp.Targets = []*Target{{
TargetId: "eu.proxy.netbird.io",
TargetType: TargetTypeCluster,
Protocol: "tcp",
Enabled: true,
}}
assert.ErrorContains(t, rp.Validate(), "HTTP")
}
@@ -20,6 +20,20 @@ type KeyPair struct {
type Claims struct {
jwt.RegisteredClaims
Method auth.Method `json:"method"`
// Email is the calling user's email address. Carried so the
// proxy can stamp identity on upstream requests (e.g.
// x-litellm-end-user-id) without an extra management
// round-trip on every cookie-bearing request.
Email string `json:"email,omitempty"`
// Groups carries the user's group IDs so the proxy can stamp them
// onto upstream requests (X-NetBird-Groups) from the cookie path
// without an extra management round-trip.
Groups []string `json:"groups,omitempty"`
// GroupNames carries the human-readable display names for the ids
// in Groups, ordered identically (positional pairing). Slice may be
// shorter than Groups for tokens minted before names were
// resolvable; the consumer falls back to ids for missing positions.
GroupNames []string `json:"group_names,omitempty"`
}
func GenerateKeyPair() (*KeyPair, error) {
@@ -34,7 +48,13 @@ func GenerateKeyPair() (*KeyPair, error) {
}, nil
}
func SignToken(privKeyB64, userID, domain string, method auth.Method, expiration time.Duration) (string, error) {
// SignToken mints a session JWT for the given user and domain. email,
// groups, and groupNames, when non-empty, are embedded so the proxy can
// authorise and stamp identity for policy-aware middlewares without a
// management round-trip on every cookie-bearing request. groupNames
// pairs positionally with groups; pass nil when names couldn't be
// resolved.
func SignToken(privKeyB64, userID, email, domain string, method auth.Method, groups, groupNames []string, expiration time.Duration) (string, error) {
privKeyBytes, err := base64.StdEncoding.DecodeString(privKeyB64)
if err != nil {
return "", fmt.Errorf("decode private key: %w", err)
@@ -56,7 +76,10 @@ func SignToken(privKeyB64, userID, domain string, method auth.Method, expiration
IssuedAt: jwt.NewNumericDate(now),
NotBefore: jwt.NewNumericDate(now),
},
Method: method,
Method: method,
Email: email,
Groups: append([]string(nil), groups...),
GroupNames: append([]string(nil), groupNames...),
}
token := jwt.NewWithClaims(jwt.SigningMethodEdDSA, claims)