[client] Add local Prometheus metrics endpoint (#6689)

client/server/mdm.go

@@ -233,6 +233,24 @@ func conflictString(key, got string) conflictCheck {
 	}
 }
 
+// conflictStringPtr is conflictString for optional proto fields, where an
+// explicit empty value is still a request to change the setting. If p is
+// nil the field is treated as matching (no override requested); otherwise
+// the check returns true only when the policy contains the key and its
+// value equals *p.
+func conflictStringPtr(key string, p *string) conflictCheck {
+	return conflictCheck{
+		key: key,
+		check: func(pol *mdm.Policy) bool {
+			if p == nil {
+				return true
+			}
+			want, ok := pol.GetString(key)
+			return ok && want == *p
+		},
+	}
+}
+
 // conflictInt64 builds a conflictCheck for an integer MDM key. If p is
 // nil the field is treated as matching; otherwise the check returns
 // true only when the policy contains the key and its int value equals *p.
@@ -301,6 +319,8 @@ func mdmManagedFieldConflicts(msg *proto.SetConfigRequest, policy *mdm.Policy) [
 		conflictBool(mdm.KeyDisableServerRoutes, msg.DisableServerRoutes),
 		conflictBool(mdm.KeyBlockInbound, msg.BlockInbound),
 		conflictInt64(mdm.KeyWireguardPort, msg.WireguardPort),
+		conflictBool(mdm.KeyEnableLocalMetrics, msg.EnableLocalMetrics),
+		conflictStringPtr(mdm.KeyLocalMetricsAddress, msg.LocalMetricsAddress),
 	})
 }
 
@@ -346,7 +366,9 @@ func setConfigRequestHasConfigOverrides(msg *proto.SetConfigRequest) bool {
 		msg.EnableSSHLocalPortForwarding != nil ||
 		msg.EnableSSHRemotePortForwarding != nil ||
 		msg.DisableSSHAuth != nil ||
-		msg.SshJWTCacheTTL != nil
+		msg.SshJWTCacheTTL != nil ||
+		msg.EnableLocalMetrics != nil ||
+		msg.LocalMetricsAddress != nil
 }
 
 // loginRequestHasConfigOverrides reports whether the LoginRequest
@@ -381,7 +403,9 @@ func loginRequestHasConfigOverrides(msg *proto.LoginRequest) bool {
 		msg.BlockLanAccess != nil ||
 		msg.DisableNotifications != nil ||
 		len(msg.DnsLabels) > 0 || msg.CleanDNSLabels ||
-		msg.BlockInbound != nil
+		msg.BlockInbound != nil ||
+		msg.EnableLocalMetrics != nil ||
+		msg.LocalMetricsAddress != nil
 }
 
 // loginRequestMDMConflicts mirrors mdmManagedFieldConflicts but for the
@@ -422,6 +446,8 @@ func loginRequestMDMConflicts(msg *proto.LoginRequest, policy *mdm.Policy) []str
 		conflictBool(mdm.KeyDisableServerRoutes, msg.DisableServerRoutes),
 		conflictBool(mdm.KeyBlockInbound, msg.BlockInbound),
 		conflictInt64(mdm.KeyWireguardPort, msg.WireguardPort),
+		conflictBool(mdm.KeyEnableLocalMetrics, msg.EnableLocalMetrics),
+		conflictStringPtr(mdm.KeyLocalMetricsAddress, msg.LocalMetricsAddress),
 	})
 }
 

client/server/server.go

@@ -23,6 +23,9 @@ import (
 
 	"github.com/netbirdio/netbird/client/internal/auth"
 	"github.com/netbirdio/netbird/client/internal/expose"
+	"github.com/prometheus/client_golang/prometheus"
+
+	"github.com/netbirdio/netbird/client/internal/localmetrics"
 	"github.com/netbirdio/netbird/client/internal/profilemanager"
 	sleephandler "github.com/netbirdio/netbird/client/internal/sleep/handler"
 	"github.com/netbirdio/netbird/client/mdm"
@@ -108,6 +111,7 @@ type Server struct {
 
 	statusRecorder *peer.Status
 	sessionWatcher *internal.SessionWatcher
+	localMetrics   *localmetrics.Manager
 
 	probeThrottle       *probeThrottle
 	persistSyncResponse bool
@@ -171,9 +175,28 @@ func New(ctx context.Context, logFile string, configFile string, profilesDisable
 	s.sleepHandler = sleephandler.New(agent)
 	s.startSleepDetector()
 
+	s.localMetrics = localmetrics.NewManager(ctx, s.statusRecorder, s.clientMetricsGatherer)
+
 	return s
 }
 
+// clientMetricsGatherer returns the Prometheus gatherer of the running
+// engine's client metrics, or nil when no engine is running.
+func (s *Server) clientMetricsGatherer() prometheus.Gatherer {
+	s.mutex.Lock()
+	connectClient := s.connectClient
+	s.mutex.Unlock()
+
+	if connectClient == nil {
+		return nil
+	}
+	engine := connectClient.Engine()
+	if engine == nil {
+		return nil
+	}
+	return engine.GetClientMetrics().PrometheusGatherer()
+}
+
 func (s *Server) Start() error {
 	s.mutex.Lock()
 	defer s.mutex.Unlock()
@@ -254,6 +277,7 @@ func (s *Server) Start() error {
 
 	s.statusRecorder.UpdateManagementAddress(config.ManagementURL.String())
 	s.statusRecorder.UpdateRosenpass(config.RosenpassEnabled, config.RosenpassPermissive)
+	s.localMetrics.Reconcile(config.LocalMetricsEnabled, config.LocalMetricsAddress)
 
 	if s.sessionWatcher == nil {
 		s.sessionWatcher = internal.NewSessionWatcher(s.rootCtx, s.statusRecorder)
@@ -477,11 +501,18 @@ func (s *Server) SetConfig(callerCtx context.Context, msg *proto.SetConfigReques
 		return nil, err
 	}
 
-	if _, err := profilemanager.UpdateConfig(config); err != nil {
+	updatedConf, err := profilemanager.UpdateConfig(config)
+	if err != nil {
 		log.Errorf("failed to update profile config: %v", err)
 		return nil, fmt.Errorf("failed to update profile config: %w", err)
 	}
 
+	if activeProf, err := s.profileManager.GetActiveProfileState(); err == nil {
+		if activePath, err := activeProf.FilePath(); err == nil && activePath == config.ConfigPath {
+			s.localMetrics.Reconcile(updatedConf.LocalMetricsEnabled, updatedConf.LocalMetricsAddress)
+		}
+	}
+
 	return &proto.SetConfigResponse{}, nil
 }
 
@@ -551,6 +582,8 @@ func (s *Server) setConfigInputFromRequest(msg *proto.SetConfigRequest) (profile
 
 	config.RosenpassEnabled = msg.RosenpassEnabled
 	config.RosenpassPermissive = msg.RosenpassPermissive
+	config.LocalMetricsEnabled = msg.EnableLocalMetrics
+	config.LocalMetricsAddress = msg.LocalMetricsAddress
 	config.DisableAutoConnect = msg.DisableAutoConnect
 	config.ServerSSHAllowed = msg.ServerSSHAllowed
 	config.NetworkMonitor = msg.NetworkMonitor
@@ -657,6 +690,8 @@ func (s *Server) Login(callerCtx context.Context, msg *proto.LoginRequest) (*pro
 	s.config = config
 	s.mutex.Unlock()
 
+	s.localMetrics.Reconcile(config.LocalMetricsEnabled, config.LocalMetricsAddress)
+
 	// A probe that errors leaves the login undecided: Management unreachable, a
 	// restart mid-request, an internal error. Those are returned for the caller
 	// to retry, because turning them into an SSO prompt asks the user to solve
@@ -1007,6 +1042,7 @@ func (s *Server) Up(callerCtx context.Context, msg *proto.UpRequest) (*proto.UpR
 
 	s.statusRecorder.UpdateManagementAddress(s.config.ManagementURL.String())
 	s.statusRecorder.UpdateRosenpass(s.config.RosenpassEnabled, s.config.RosenpassPermissive)
+	s.localMetrics.Reconcile(s.config.LocalMetricsEnabled, s.config.LocalMetricsAddress)
 
 	s.clientRunning = true
 	s.clientRunningChan = make(chan struct{})
@@ -1184,6 +1220,7 @@ func (s *Server) SwitchProfile(callerCtx context.Context, msg *proto.SwitchProfi
 	}
 
 	s.config = config
+	s.localMetrics.Reconcile(config.LocalMetricsEnabled, config.LocalMetricsAddress)
 
 	if msg != nil && msg.ProfileName != nil {
 		s.publishProfileListChanged(*msg.ProfileName)

client/server/setconfig_mdm_test.go

@@ -136,6 +136,51 @@ func TestSetConfig_MDMReject_MultipleFields(t *testing.T) {
 	}, v.GetFields())
 }
 
+func TestSetConfig_MDMReject_LocalMetrics(t *testing.T) {
+	withMDMPolicy(t, mdm.NewPolicy(map[string]any{
+		mdm.KeyEnableLocalMetrics:  true,
+		mdm.KeyLocalMetricsAddress: "127.0.0.1:9191",
+	}))
+
+	s, ctx, profName, username, _ := setupServerWithProfile(t)
+
+	enabled := false
+	addr := "0.0.0.0:9999"
+	_, err := s.SetConfig(ctx, &proto.SetConfigRequest{
+		ProfileName:         profName,
+		Username:            username,
+		EnableLocalMetrics:  &enabled,
+		LocalMetricsAddress: &addr,
+	})
+
+	v := extractViolation(t, err)
+	assert.ElementsMatch(t, []string{
+		mdm.KeyEnableLocalMetrics,
+		mdm.KeyLocalMetricsAddress,
+	}, v.GetFields())
+}
+
+// An explicitly empty address still changes the effective listen address
+// (the manager falls back to the default), so presence must be honored
+// rather than collapsed to "field not set".
+func TestSetConfig_MDMReject_LocalMetricsEmptyAddress(t *testing.T) {
+	withMDMPolicy(t, mdm.NewPolicy(map[string]any{
+		mdm.KeyLocalMetricsAddress: "127.0.0.1:9999",
+	}))
+
+	s, ctx, profName, username, _ := setupServerWithProfile(t)
+
+	addr := ""
+	_, err := s.SetConfig(ctx, &proto.SetConfigRequest{
+		ProfileName:         profName,
+		Username:            username,
+		LocalMetricsAddress: &addr,
+	})
+
+	v := extractViolation(t, err)
+	assert.ElementsMatch(t, []string{mdm.KeyLocalMetricsAddress}, v.GetFields())
+}
+
 func TestSetConfig_MDMReject_AllOrNothing(t *testing.T) {
 	// MDM enforces ManagementURL only; user request touches both the
 	// enforced field AND a non-enforced field (RosenpassEnabled).

client/server/setconfig_test.go

@@ -76,6 +76,8 @@ func TestSetConfig_AllFieldsSaved(t *testing.T) {
 	disableIPv6 := true
 	mtu := int64(1280)
 	sshJWTCacheTTL := int32(300)
+	enableLocalMetrics := true
+	localMetricsAddress := "127.0.0.1:9292"
 
 	req := &proto.SetConfigRequest{
 		ProfileName:          profName,
@@ -107,6 +109,8 @@ func TestSetConfig_AllFieldsSaved(t *testing.T) {
 		DnsRouteInterval:     durationpb.New(2 * time.Minute),
 		Mtu:                  &mtu,
 		SshJWTCacheTTL:       &sshJWTCacheTTL,
+		EnableLocalMetrics:   &enableLocalMetrics,
+		LocalMetricsAddress:  &localMetricsAddress,
 	}
 
 	_, err = s.SetConfig(ctx, req)
@@ -153,6 +157,8 @@ func TestSetConfig_AllFieldsSaved(t *testing.T) {
 	require.Equal(t, uint16(mtu), cfg.MTU)
 	require.NotNil(t, cfg.SSHJWTCacheTTL)
 	require.Equal(t, int(sshJWTCacheTTL), *cfg.SSHJWTCacheTTL)
+	require.Equal(t, enableLocalMetrics, cfg.LocalMetricsEnabled)
+	require.Equal(t, localMetricsAddress, cfg.LocalMetricsAddress)
 
 	verifyAllFieldsCovered(t, req)
 }
@@ -205,6 +211,8 @@ func verifyAllFieldsCovered(t *testing.T, req *proto.SetConfigRequest) {
 		"EnableSSHRemotePortForwarding": true,
 		"DisableSSHAuth":                true,
 		"SshJWTCacheTTL":                true,
+		"EnableLocalMetrics":            true,
+		"LocalMetricsAddress":           true,
 	}
 
 	val := reflect.ValueOf(req).Elem()
@@ -264,6 +272,8 @@ func TestCLIFlags_MappedToSetConfig(t *testing.T) {
 		"enable-ssh-remote-port-forwarding": "EnableSSHRemotePortForwarding",
 		"disable-ssh-auth":                  "DisableSSHAuth",
 		"ssh-jwt-cache-ttl":                 "SshJWTCacheTTL",
+		"enable-local-metrics":              "EnableLocalMetrics",
+		"local-metrics-address":             "LocalMetricsAddress",
 	}
 
 	// SetConfigRequest fields that don't have CLI flags (settable only via UI or other means).

client/server/ssh_gate.go

@@ -14,6 +14,7 @@ import (
 
 	"github.com/netbirdio/netbird/client/internal/daemonaddr"
 	"github.com/netbirdio/netbird/client/internal/ipcauth"
+	"github.com/netbirdio/netbird/client/internal/localmetrics"
 	"github.com/netbirdio/netbird/client/internal/profilemanager"
 	"github.com/netbirdio/netbird/client/proto"
 	"github.com/netbirdio/netbird/util"
@@ -30,6 +31,8 @@ import (
 //     management identity hands SSH authorization decisions, including which
 //     keys and users are accepted, to whoever controls that identity. Changing
 //     the management URL and deregistering the peer are both ways to do that.
+//   - Binding the local metrics endpoint to a non-loopback address publishes
+//     peer names and connectivity state to the network without authentication.
 //
 // Everything else stays unauthenticated, so this is not an authorization model:
 // it only refuses the changes that would let a local user become root. A caller
@@ -39,27 +42,33 @@ import (
 // user-to-root boundary. Fields are nil or empty when the request leaves them
 // untouched.
 type privilegedConfigChange struct {
-	managementURL    string
-	serverSSHAllowed *bool
-	enableSSHRoot    *bool
-	disableSSHAuth   *bool
+	managementURL       string
+	serverSSHAllowed    *bool
+	enableSSHRoot       *bool
+	disableSSHAuth      *bool
+	enableLocalMetrics  *bool
+	localMetricsAddress *string
 }
 
 func privilegedChangeFromSetConfig(msg *proto.SetConfigRequest) privilegedConfigChange {
 	return privilegedConfigChange{
-		managementURL:    msg.GetManagementUrl(),
-		serverSSHAllowed: msg.ServerSSHAllowed,
-		enableSSHRoot:    msg.EnableSSHRoot,
-		disableSSHAuth:   msg.DisableSSHAuth,
+		managementURL:       msg.GetManagementUrl(),
+		serverSSHAllowed:    msg.ServerSSHAllowed,
+		enableSSHRoot:       msg.EnableSSHRoot,
+		disableSSHAuth:      msg.DisableSSHAuth,
+		enableLocalMetrics:  msg.EnableLocalMetrics,
+		localMetricsAddress: msg.LocalMetricsAddress,
 	}
 }
 
 func privilegedChangeFromLogin(msg *proto.LoginRequest) privilegedConfigChange {
 	return privilegedConfigChange{
-		managementURL:    msg.GetManagementUrl(),
-		serverSSHAllowed: msg.ServerSSHAllowed,
-		enableSSHRoot:    msg.EnableSSHRoot,
-		disableSSHAuth:   msg.DisableSSHAuth,
+		managementURL:       msg.GetManagementUrl(),
+		serverSSHAllowed:    msg.ServerSSHAllowed,
+		enableSSHRoot:       msg.EnableSSHRoot,
+		disableSSHAuth:      msg.DisableSSHAuth,
+		enableLocalMetrics:  msg.EnableLocalMetrics,
+		localMetricsAddress: msg.LocalMetricsAddress,
 	}
 }
 
@@ -83,6 +92,12 @@ func requirePrivilegeForConfigChange(ctx context.Context, stored *profilemanager
 		return denyPrivileged(ctx, "enabling the NetBird SSH server", ipcauth.UpCommand("--allow-server-ssh"))
 	}
 
+	if addr, exposes := exposesLocalMetrics(stored, change); exposes {
+		return denyPrivileged(ctx,
+			"exposing the local metrics endpoint on a non-loopback address",
+			ipcauth.UpCommand("--enable-local-metrics --local-metrics-address "+addr))
+	}
+
 	// Only guard the management binding while the SSH server is enabled: that is
 	// when the management identity decides who may open a shell here.
 	if !sshServerEnabled(stored) {
@@ -245,6 +260,48 @@ func sshServerCurrentlyAllowed(cfg *profilemanager.Config) *bool {
 	return &enabled
 }
 
+// exposesLocalMetrics reports whether the change would leave the metrics
+// endpoint enabled on an address that is not confirmed loopback, and returns
+// that address. A request that restates the stored state is not a change, so a
+// settings form resubmitted after an administrator opened the endpoint is not
+// refused.
+func exposesLocalMetrics(stored *profilemanager.Config, change privilegedConfigChange) (string, bool) {
+	storedEnabled, storedAddr := storedLocalMetrics(stored)
+
+	enabled := storedEnabled
+	if change.enableLocalMetrics != nil {
+		enabled = *change.enableLocalMetrics
+	}
+	addr := storedAddr
+	if change.localMetricsAddress != nil {
+		addr = metricsAddrOrDefault(*change.localMetricsAddress)
+	}
+
+	if !enabled || localmetrics.IsLoopback(addr) {
+		return "", false
+	}
+	if storedEnabled && storedAddr == addr {
+		return "", false
+	}
+	return addr, true
+}
+
+// storedLocalMetrics reads the metrics settings from the stored config,
+// tolerating a config that does not exist yet.
+func storedLocalMetrics(cfg *profilemanager.Config) (bool, string) {
+	if cfg == nil {
+		return false, localmetrics.DefaultListenAddress
+	}
+	return cfg.LocalMetricsEnabled, metricsAddrOrDefault(cfg.LocalMetricsAddress)
+}
+
+func metricsAddrOrDefault(addr string) string {
+	if addr == "" {
+		return localmetrics.DefaultListenAddress
+	}
+	return addr
+}
+
 // sameManagementURL reports whether requested addresses the same management
 // server as stored, comparing scheme, host and effective port so that an
 // equivalent spelling ("https://api.netbird.io" for a stored

client/server/ssh_gate_test.go

@@ -61,6 +61,8 @@ func noIdentityCtx() context.Context { return context.Background() }
 
 func boolPtr(v bool) *bool { return &v }
 
+func strPtr(v string) *string { return &v }
+
 func mustURL(t *testing.T, raw string) *url.URL {
 	t.Helper()
 	u, err := url.Parse(raw)
@@ -194,6 +196,102 @@ func TestRequirePrivilegeForConfigChange_SSHFlags(t *testing.T) {
 	}
 }
 
+func TestRequirePrivilegeForConfigChange_LocalMetrics(t *testing.T) {
+	exposed := &profilemanager.Config{LocalMetricsEnabled: true, LocalMetricsAddress: "0.0.0.0:9191"}
+
+	tests := []struct {
+		name       string
+		stored     *profilemanager.Config
+		change     privilegedConfigChange
+		privileged bool
+		wantDeny   bool
+	}{
+		{
+			name:     "binding a non-loopback address unprivileged is refused",
+			stored:   &profilemanager.Config{},
+			change:   privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("0.0.0.0:9191")},
+			wantDeny: true,
+		},
+		{
+			name:       "binding a non-loopback address as root is allowed",
+			stored:     &profilemanager.Config{},
+			change:     privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("0.0.0.0:9191")},
+			privileged: true,
+		},
+		{
+			name:   "enabling on the default loopback address is not guarded",
+			stored: &profilemanager.Config{},
+			change: privilegedConfigChange{enableLocalMetrics: boolPtr(true)},
+		},
+		{
+			name:   "enabling on an explicit loopback address is not guarded",
+			stored: &profilemanager.Config{},
+			change: privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("127.0.0.1:9999")},
+		},
+		{
+			name:   "enabling on the IPv6 loopback address is not guarded",
+			stored: &profilemanager.Config{},
+			change: privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("[::1]:9191")},
+		},
+		{
+			// The address alone does nothing while the endpoint stays off.
+			name:   "a non-loopback address without enabling is not guarded",
+			stored: &profilemanager.Config{},
+			change: privilegedConfigChange{localMetricsAddress: strPtr("0.0.0.0:9191")},
+		},
+		{
+			name:     "widening an already enabled loopback endpoint is refused",
+			stored:   &profilemanager.Config{LocalMetricsEnabled: true, LocalMetricsAddress: "127.0.0.1:9191"},
+			change:   privilegedConfigChange{localMetricsAddress: strPtr("0.0.0.0:9191")},
+			wantDeny: true,
+		},
+		{
+			name:   "restating an already exposed endpoint is not a change",
+			stored: exposed,
+			change: privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("0.0.0.0:9191")},
+		},
+		{
+			name:   "turning an exposed endpoint off is not guarded",
+			stored: exposed,
+			change: privilegedConfigChange{enableLocalMetrics: boolPtr(false)},
+		},
+		{
+			name:     "re-enabling an exposed endpoint that was turned off is refused",
+			stored:   &profilemanager.Config{LocalMetricsEnabled: false, LocalMetricsAddress: "0.0.0.0:9191"},
+			change:   privilegedConfigChange{enableLocalMetrics: boolPtr(true)},
+			wantDeny: true,
+		},
+		{
+			// Fail closed: an address that cannot be parsed is not confirmed loopback.
+			name:     "an unparseable address is refused",
+			stored:   &profilemanager.Config{},
+			change:   privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("not-an-address")},
+			wantDeny: true,
+		},
+		{
+			name:     "a profile with no config yet counts as off, so exposing is refused",
+			stored:   nil,
+			change:   privilegedConfigChange{enableLocalMetrics: boolPtr(true), localMetricsAddress: strPtr("0.0.0.0:9191")},
+			wantDeny: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			ctx := userCtx()
+			if tt.privileged {
+				ctx = rootCtx()
+			}
+			err := requirePrivilegeForConfigChange(ctx, tt.stored, tt.change)
+			if tt.wantDeny {
+				assertDenied(t, err)
+				return
+			}
+			assertAllowed(t, err)
+		})
+	}
+}
+
 func TestRequirePrivilegeForConfigChange_ManagementURL(t *testing.T) {
 	sshOn := func(raw string) *profilemanager.Config {
 		return &profilemanager.Config{ServerSSHAllowed: boolPtr(true), ManagementURL: mustURL(t, raw)}