RSDK-11248 Disable restart checking logic with new agents and update restart_status endpoint

robot/web/web.go

@@ -971,14 +971,23 @@ func (svc *webService) Stats() any {
 // RestartStatusResponse is the JSON response of the `restart_status` HTTP
 // endpoint.
 type RestartStatusResponse struct {
-	// RestartAllowed represents whether this instance of the viam-server can be
+	// RestartAllowed represents whether this instance of the viamserver can be
 	// safely restarted.
 	RestartAllowed bool `json:"restart_allowed"`
+	// DoesNotHandleNeedsRestart represents whether this instance of the viamserver does
+	// not check for the need to restart against app itself and, thus, needs agent to do so.
+	// Newer versions of viamserver (>= v0.9x.0) will report true for this value, while
+	// older versions won't report it at all, and agent should let viamserver handle
+	// NeedsRestart logic.
+	DoesNotHandleNeedsRestart bool `json:"does_not_handle_needs_restart,omitempty"`
 }
 
 // Handles the `/restart_status` endpoint.
 func (svc *webService) handleRestartStatus(w http.ResponseWriter, r *http.Request) {
-	response := RestartStatusResponse{RestartAllowed: svc.r.RestartAllowed()}
+	response := RestartStatusResponse{
+		RestartAllowed:            svc.r.RestartAllowed(),
+		DoesNotHandleNeedsRestart: true,
+	}
 
 	w.Header().Set("Content-Type", "application/json")
 	// Only log errors from encoding here. A failure to encode should never

utils/env.go

@@ -79,6 +79,16 @@ const (
 
 	// GetImagesInStreamServerEnvVar is the environment variable that enables the GetImages feature flag in stream server.
 	GetImagesInStreamServerEnvVar = "VIAM_GET_IMAGES_IN_STREAM_SERVER"
+
+	// ViamAgentHandlesNeedsRestartChecking is the environment variable that viam-agent will
+	// set before starting viam-server to indicate that agent is a new enough version to
+	// have its own background loop that runs NeedsRestart against app.viam.com to determine
+	// if the system needs a restart. MUST be kept in line with the equivalent value in the
+	// agent repo.
+	//
+	// TODO(RSDK-12057): Remove sensitivity to this environment variable once we fully
+	// remove all NeedsRestart checking logic from viam-server.
+	ViamAgentHandlesNeedsRestartChecking = "VIAM_AGENT_HANDLES_NEEDS_RESTART_CHECKING"
 )
 
 // EnvTrueValues contains strings that we interpret as boolean true in env vars.

web/cmd/droid/droid.go

@@ -19,7 +19,11 @@ var logger = logging.NewDebugLogger("droid-entrypoint")
 
 // DroidStopHook used by android harness to stop the RDK.
 func DroidStopHook() { //nolint:revive
-	server.ForceRestart = true
+	// NOTE(benjirewis): In RSDK-11248, we removed the restart checking logic from
+	// viam-server and put it in viam-agent. This method used to set a flag used by restart
+	// checking logic, but since that no longer exists and we have no users of this droid
+	// code, we no longer support this method.
+	logger.Error("DroidStopHook is no longer supported")
 }
 
 // MainEntry is called by our android app to start the RDK.

web/server/entrypoint.go

@@ -529,7 +529,10 @@ func (s *robotServer) serveWeb(ctx context.Context, cfg *config.Config) (err err
 	// This functionality is tested in `TestLogPropagation` in `local_robot_test.go`.
 	config.UpdateLoggerRegistryFromConfig(s.registry, fullProcessedConfig, s.logger)
 
-	if fullProcessedConfig.Cloud != nil {
+	// Only start cloud restart checker if cloud config is non-nil, and viam-agent is not
+	// handling restart checking for us (relevant environment variable is unset).
+	if fullProcessedConfig.Cloud != nil && os.Getenv(rutils.ViamAgentHandlesNeedsRestartChecking) == "" {
+		s.logger.CInfo(ctx, "Agent does not handle checking needs restart functionality; will handle in server")
 		cloudRestartCheckerActive = make(chan struct{})
 		utils.PanicCapturingGo(func() {
 			defer close(cloudRestartCheckerActive)

web/server/restart_checker.go

@@ -27,13 +27,7 @@ type needsRestartCheckerGRPC struct {
 	client rpc.ClientConn
 }
 
-// ForceRestart lets random other parts of the app request an exit.
-var ForceRestart bool
-
 func (c *needsRestartCheckerGRPC) needsRestart(ctx context.Context) (bool, time.Duration, error) {
-	if ForceRestart {
-		return true, minNeedsRestartCheckInterval, nil
-	}
 	service := apppb.NewRobotServiceClient(c.client)
 	res, err := service.NeedsRestart(ctx, &apppb.NeedsRestartRequest{Id: c.cfg.ID})
 	if err != nil {