[cmd/opampsupervisor] Handle OpAMP connection settings

.chloggen/supervisor-accepts-conn.yaml

@@ -0,0 +1,27 @@
+# Use this changelog template to create an entry for release notes.
+
+# One of 'breaking', 'deprecation', 'new_component', 'enhancement', 'bug_fix'
+change_type: enhancement
+
+# The name of the component, or a single word describing the area of concern, (e.g. filelogreceiver)
+component: cmd/opampsupervisor
+
+# A brief description of the change.  Surround your text with quotes ("") if it needs to start with a backtick (`).
+note: Handle OpAMP connection settings.
+
+# Mandatory: One or more tracking issues related to the change. You can use the PR number here if no issue exists.
+issues: [21043]
+
+# (Optional) One or more lines of additional information to render under the primary note.
+# These lines will be padded with 2 spaces and then inserted directly into the document.
+# Use pipe (|) for multiline entries.
+subtext:
+
+# If your change doesn't affect end users or the exported elements of any package,
+# you should instead start your pull request title with [chore] or use the "Skip Changelog" label.
+# Optional: The change log or logs in which this entry should be included.
+# e.g. '[user]' or '[user, api]'
+# Include 'user' if the change is relevant to end users.
+# Include 'api' if there is a change to a library API.
+# Default: '[user]'
+change_logs: []

cmd/opampsupervisor/e2e_test.go

@@ -81,13 +81,13 @@ func newOpAMPServer(t *testing.T, connectingCallback onConnectingFuncFactory, ca
 	connectedChan := make(chan bool)
 	s := server.New(testLogger{t: t})
 	onConnectedFunc := callbacks.OnConnectedFunc
-	callbacks.OnConnectedFunc = func(conn types.Connection) {
+	callbacks.OnConnectedFunc = func(ctx context.Context, conn types.Connection) {
+		if onConnectedFunc != nil {
+			onConnectedFunc(ctx, conn)
+		}
 		agentConn.Store(conn)
 		isAgentConnected.Store(true)
 		connectedChan <- true
-		if onConnectedFunc != nil {
-			onConnectedFunc(conn)
-		}
 	}
 	onConnectionCloseFunc := callbacks.OnConnectionCloseFunc
 	callbacks.OnConnectionCloseFunc = func(conn types.Connection) {
@@ -133,7 +133,9 @@ func newOpAMPServer(t *testing.T, connectingCallback onConnectingFuncFactory, ca
 
 func newSupervisor(t *testing.T, configType string, extraConfigData map[string]string) *supervisor.Supervisor {
 	cfgFile := getSupervisorConfig(t, configType, extraConfigData)
-	s, err := supervisor.NewSupervisor(zap.NewNop(), cfgFile.Name())
+	logger, err := zap.NewDevelopment()
+	require.NoError(t, err)
+	s, err := supervisor.NewSupervisor(logger, cfgFile.Name())
 	require.NoError(t, err)
 
 	return s
@@ -175,7 +177,7 @@ func TestSupervisorStartsCollectorWithRemoteConfig(t *testing.T) {
 		t,
 		defaultConnectingHandler,
 		server.ConnectionCallbacksStruct{
-			OnMessageFunc: func(_ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
+			OnMessageFunc: func(_ context.Context, _ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
 				if message.EffectiveConfig != nil {
 					config := message.EffectiveConfig.ConfigMap.ConfigMap[""]
 					if config != nil {
@@ -235,7 +237,7 @@ func TestSupervisorRestartsCollectorAfterBadConfig(t *testing.T) {
 		t,
 		defaultConnectingHandler,
 		server.ConnectionCallbacksStruct{
-			OnMessageFunc: func(_ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
+			OnMessageFunc: func(_ context.Context, _ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
 				if message.Health != nil {
 					healthReport.Store(message.Health)
 				}
@@ -319,7 +321,7 @@ func TestSupervisorConfiguresCapabilities(t *testing.T) {
 		t,
 		defaultConnectingHandler,
 		server.ConnectionCallbacksStruct{
-			OnMessageFunc: func(_ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
+			OnMessageFunc: func(_ context.Context, _ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
 				capabilities.Store(message.Capabilities)
 
 				return &protobufs.ServerToAgent{}
@@ -372,7 +374,7 @@ func TestSupervisorBootstrapsCollector(t *testing.T) {
 		t,
 		defaultConnectingHandler,
 		server.ConnectionCallbacksStruct{
-			OnMessageFunc: func(_ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
+			OnMessageFunc: func(_ context.Context, _ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
 				if message.AgentDescription != nil {
 					agentDescription.Store(message.AgentDescription)
 				}
@@ -473,3 +475,48 @@ func waitForSupervisorConnection(connection chan bool, connected bool) {
 		}
 	}
 }
+
+func TestSupervisorOpAMPConnectionSettings(t *testing.T) {
+	var connectedToNewServer atomic.Bool
+	initialServer := newOpAMPServer(
+		t,
+		defaultConnectingHandler,
+		server.ConnectionCallbacksStruct{})
+
+	s := newSupervisor(t, "accepts_conn", map[string]string{"url": initialServer.addr})
+	defer s.Shutdown()
+
+	waitForSupervisorConnection(initialServer.supervisorConnected, true)
+
+	newServer := newOpAMPServer(
+		t,
+		defaultConnectingHandler,
+		server.ConnectionCallbacksStruct{
+			OnConnectedFunc: func(_ context.Context, _ types.Connection) {
+				connectedToNewServer.Store(true)
+			},
+			OnMessageFunc: func(_ context.Context, _ types.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
+				return &protobufs.ServerToAgent{}
+			},
+		})
+
+	initialServer.sendToSupervisor(&protobufs.ServerToAgent{
+		ConnectionSettings: &protobufs.ConnectionSettingsOffers{
+			Opamp: &protobufs.OpAMPConnectionSettings{
+				DestinationEndpoint: "ws://" + newServer.addr + "/v1/opamp",
+				Headers: &protobufs.Headers{
+					Headers: []*protobufs.Header{
+						{
+							Key:   "x-foo",
+							Value: "bar",
+						},
+					},
+				},
+			},
+		},
+	})
+
+	require.Eventually(t, func() bool {
+		return connectedToNewServer.Load() == true
+	}, 10*time.Second, 500*time.Millisecond, "Collector did not connect to new OpAMP server")
+}

cmd/opampsupervisor/go.mod

@@ -9,8 +9,9 @@ require (
 	github.com/knadh/koanf/providers/rawbytes v0.1.0
 	github.com/knadh/koanf/v2 v2.0.1
 	github.com/oklog/ulid/v2 v2.1.0
-	github.com/open-telemetry/opamp-go v0.11.0
+	github.com/open-telemetry/opamp-go v0.11.1-0.20240123204604-4d07a6af062f
 	github.com/stretchr/testify v1.8.4
+	go.opentelemetry.io/collector/config/configopaque v0.93.1-0.20240124123350-9047c0e373f9
 	go.opentelemetry.io/collector/config/configtls v0.93.1-0.20240124123350-9047c0e373f9
 	go.opentelemetry.io/collector/semconv v0.93.1-0.20240124123350-9047c0e373f9
 	go.uber.org/goleak v1.3.0
@@ -27,7 +28,6 @@ require (
 	github.com/mitchellh/mapstructure v1.5.1-0.20231216201459-8508981c8b6c // indirect
 	github.com/mitchellh/reflectwalk v1.0.2 // indirect
 	github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 // indirect
-	go.opentelemetry.io/collector/config/configopaque v0.93.1-0.20240124123350-9047c0e373f9 // indirect
 	go.uber.org/multierr v1.11.0 // indirect
 	golang.org/x/net v0.17.0 // indirect
 	golang.org/x/sys v0.16.0 // indirect

cmd/opampsupervisor/supervisor/config/config.go

@@ -4,6 +4,8 @@
 package config
 
 import (
+	"net/http"
+
 	"go.opentelemetry.io/collector/config/configtls"
 )
 
@@ -16,15 +18,17 @@ type Supervisor struct {
 
 // Capabilities is the set of capabilities that the Supervisor supports.
 type Capabilities struct {
-	AcceptsRemoteConfig    *bool `mapstructure:"accepts_remote_config"`
-	ReportsEffectiveConfig *bool `mapstructure:"reports_effective_config"`
-	ReportsOwnMetrics      *bool `mapstructure:"reports_own_metrics"`
-	ReportsHealth          *bool `mapstructure:"reports_health"`
-	ReportsRemoteConfig    *bool `mapstructure:"reports_remote_config"`
+	AcceptsRemoteConfig            *bool `mapstructure:"accepts_remote_config"`
+	AcceptsOpAMPConnectionSettings *bool `mapstructure:"accepts_opamp_connection_settings"`
+	ReportsEffectiveConfig         *bool `mapstructure:"reports_effective_config"`
+	ReportsOwnMetrics              *bool `mapstructure:"reports_own_metrics"`
+	ReportsHealth                  *bool `mapstructure:"reports_health"`
+	ReportsRemoteConfig            *bool `mapstructure:"reports_remote_config"`
 }
 
 type OpAMPServer struct {
 	Endpoint   string
+	Headers    http.Header
 	TLSSetting configtls.TLSClientSetting `mapstructure:"tls,omitempty"`
 }
 

cmd/opampsupervisor/supervisor/server.go

@@ -4,6 +4,7 @@
 package supervisor
 
 import (
+	"context"
 	"net/http"
 
 	"github.com/open-telemetry/opamp-go/protobufs"
@@ -28,7 +29,7 @@ func newServerSettings(fs flattenedSettings) server.StartSettings {
 					return serverTypes.ConnectionResponse{
 						Accept: true,
 						ConnectionCallbacks: server.ConnectionCallbacksStruct{
-							OnMessageFunc: func(conn serverTypes.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
+							OnMessageFunc: func(_ context.Context, conn serverTypes.Connection, message *protobufs.AgentToServer) *protobufs.ServerToAgent {
 								if fs.onMessageFunc != nil {
 									fs.onMessageFunc(conn, message)
 								}

cmd/opampsupervisor/supervisor/supervisor.go

@@ -29,6 +29,8 @@ import (
 	"github.com/open-telemetry/opamp-go/protobufs"
 	"github.com/open-telemetry/opamp-go/server"
 	serverTypes "github.com/open-telemetry/opamp-go/server/types"
+	"go.opentelemetry.io/collector/config/configopaque"
+	"go.opentelemetry.io/collector/config/configtls"
 	semconv "go.opentelemetry.io/collector/semconv/v1.21.0"
 	"go.uber.org/zap"
 
@@ -334,6 +336,10 @@ func (s *Supervisor) Capabilities() protobufs.AgentCapabilities {
 		if c.ReportsRemoteConfig != nil && *c.ReportsRemoteConfig {
 			supportedCapabilities |= protobufs.AgentCapabilities_AgentCapabilities_ReportsRemoteConfig
 		}
+
+		if c.AcceptsOpAMPConnectionSettings != nil && *c.AcceptsOpAMPConnectionSettings {
+			supportedCapabilities |= protobufs.AgentCapabilities_AgentCapabilities_AcceptsOpAMPConnectionSettings
+		}
 	}
 	return supportedCapabilities
 }
@@ -346,8 +352,10 @@ func (s *Supervisor) startOpAMP() error {
 		return err
 	}
 
+	s.logger.Debug("Connecting to OpAMP server...", zap.String("endpoint", s.config.Server.Endpoint), zap.Any("headers", s.config.Server.Headers))
 	settings := types.StartSettings{
 		OpAMPServerURL: s.config.Server.Endpoint,
+		Header:         s.config.Server.Headers,
 		TLSConfig:      tlsConfig,
 		InstanceUid:    s.instanceID.String(),
 		Callbacks: types.CallbacksStruct{
@@ -360,12 +368,8 @@ func (s *Supervisor) startOpAMP() error {
 			OnErrorFunc: func(err *protobufs.ServerErrorResponse) {
 				s.logger.Error("Server returned an error response", zap.String("message", err.ErrorMessage))
 			},
-			OnMessageFunc: s.onMessage,
-			OnOpampConnectionSettingsFunc: func(ctx context.Context, settings *protobufs.OpAMPConnectionSettings) error {
-				// TODO: https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/21043
-				s.logger.Debug("Received ConnectionSettings request")
-				return nil
-			},
+			OnMessageFunc:                 s.onMessage,
+			OnOpampConnectionSettingsFunc: s.onOpampConnectionSettings,
 			OnOpampConnectionSettingsAcceptedFunc: func(settings *protobufs.OpAMPConnectionSettings) {
 				// TODO: https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/21043
 				s.logger.Debug("ConnectionSettings accepted")
@@ -409,6 +413,79 @@ func (s *Supervisor) startOpAMP() error {
 	return nil
 }
 
+func (s *Supervisor) stopOpAMP() error {
+	s.logger.Debug("Stopping OpAMP client...")
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+	err := s.opampClient.Stop(ctx)
+	// TODO(srikanthccv): remove context.DeadlineExceeded after https://github.com/open-telemetry/opamp-go/pull/213
+	if err != nil && !errors.Is(err, context.DeadlineExceeded) {
+		return err
+	}
+	s.logger.Debug("OpAMP client stopped.")
+	return nil
+}
+
+func (s *Supervisor) getHeadersFromSettings(protoHeaders *protobufs.Headers) http.Header {
+	headers := make(http.Header)
+	for _, header := range protoHeaders.Headers {
+		headers.Add(header.Key, header.Value)
+	}
+	return headers
+}
+
+func (s *Supervisor) onOpampConnectionSettings(_ context.Context, settings *protobufs.OpAMPConnectionSettings) error {
+	if settings == nil {
+		s.logger.Debug("Received ConnectionSettings request with nil settings")
+		return nil
+	}
+
+	newServerConfig := &config.OpAMPServer{}
+
+	if settings.DestinationEndpoint != "" {
+		newServerConfig.Endpoint = settings.DestinationEndpoint
+	}
+	if settings.Headers != nil {
+		newServerConfig.Headers = s.getHeadersFromSettings(settings.Headers)
+	}
+	if settings.Certificate != nil {
+		if len(settings.Certificate.CaPublicKey) != 0 {
+			newServerConfig.TLSSetting.CAPem = configopaque.String(settings.Certificate.CaPublicKey)
+		}
+		if len(settings.Certificate.PublicKey) != 0 {
+			newServerConfig.TLSSetting.CertPem = configopaque.String(settings.Certificate.PublicKey)
+		}
+		if len(settings.Certificate.PrivateKey) != 0 {
+			newServerConfig.TLSSetting.KeyPem = configopaque.String(settings.Certificate.PrivateKey)
+		}
+	} else {
+		newServerConfig.TLSSetting = configtls.TLSClientSetting{Insecure: true}
+	}
+
+	if err := s.stopOpAMP(); err != nil {
+		s.logger.Error("Cannot stop the OpAMP client", zap.Error(err))
+		return err
+	}
+
+	// take a copy of the current OpAMP server config
+	oldServerConfig := s.config.Server
+	// update the OpAMP server config
+	s.config.Server = newServerConfig
+
+	if err := s.startOpAMP(); err != nil {
+		s.logger.Error("Cannot connect to the OpAMP server using the new settings", zap.Error(err))
+		// revert the OpAMP server config
+		s.config.Server = oldServerConfig
+		// start the OpAMP client with the old settings
+		if err := s.startOpAMP(); err != nil {
+			s.logger.Error("Cannot reconnect to the OpAMP server after restoring old settings", zap.Error(err))
+			return err
+		}
+	}
+
+	return nil
+}
+
 // TODO: Persist instance ID. https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/21073
 func (s *Supervisor) createInstanceID() (ulid.ULID, error) {
 	entropy := ulid.Monotonic(rand.New(rand.NewSource(0)), 0)

cmd/opampsupervisor/testdata/supervisor/supervisor_accepts_conn.yaml

@@ -0,0 +1,15 @@
+server:
+  endpoint: ws://{{.url}}/v1/opamp
+  tls:
+    insecure: true
+
+capabilities:
+  reports_effective_config: true
+  reports_own_metrics: true
+  reports_health: true
+  accepts_remote_config: true
+  reports_remote_config: true
+  accepts_opamp_connection_settings: true
+
+agent:
+  executable: ../../bin/otelcontribcol_{{.goos}}_{{.goarch}}{{.extension}}