fix: #239 set up loger instance for self-hosted installation

internal/app/ptah-agent/metrics.go

@@ -82,9 +82,13 @@ func (m *MetricsAgent) ScrapeMetrics(ctx context.Context) ([]string, error) {
 }
 
 func (m *MetricsAgent) Start(ctx context.Context) error {
+	log := Logger(ctx)
+
+	log.Info("starting metrics agent")
+
 	metrics, err := m.ScrapeMetrics(ctx)
 	if err != nil {
-		log.Printf("failed to collect metrics: %v", err)
+		log.Info("failed to collect metrics", "error", err)
 	}
 
 	err = m.client.SendMetrics(ctx, metrics)
@@ -107,14 +111,14 @@ func (m *MetricsAgent) Start(ctx context.Context) error {
 
 				metrics, err := m.ScrapeMetrics(ctx)
 				if err != nil {
-					log.Printf("failed to collect metrics: %v", err)
+					log.Error("failed to collect metrics", "error", err)
 
 					continue
 				}
 
 				err = m.client.SendMetrics(ctx, metrics)
 				if err != nil {
-					log.Printf("failed to send metrics: %v", err)
+					log.Error("failed to send metrics", "error", err)
 				}
 			}
 		}

internal/app/ptah-agent/ptah_client.go

@@ -37,7 +37,9 @@ func New(version string, baseUrl string, ptahToken string, rootDir string) (*Age
 	// Create a background context for API version negotiation
 	ctx := context.Background()
 
+	slog.Debug("negotiating API version with Docker")
 	docker.NegotiateAPIVersion(ctx)
+	slog.Debug("negotiated API version with Docker", "version", docker.ClientVersion())
 
 	caddy := caddyClient.New("http://127.0.0.1:2019", http.DefaultClient)
 
@@ -172,6 +174,8 @@ func (a *Agent) Start(ctx context.Context) error {
 	ctx = WithLogger(ctx, slog.New(slog.NewTextHandler(os.Stdout, &slog.HandlerOptions{Level: level})))
 	log := Logger(ctx)
 
+	log.Info("starting ptah agent", "version", a.Version)
+
 	a.cancel = cancel
 
 	defer a.safeClient.Close()
@@ -279,6 +283,19 @@ func (a *Agent) getNextTask(ctx context.Context) (taskId int, task interface{},
 }
 
 func (a *Agent) ExecTasks(ctx context.Context, jsonFilePath string) error {
+	// FIXME: de-duplicate logger instantiation with Start()
+	ptahDebug := os.Getenv("PTAH_DEBUG")
+
+	var level slog.Level
+	if ptahDebug == "true" {
+		level = slog.LevelDebug
+	} else {
+		level = slog.LevelInfo
+	}
+
+	ctx = WithLogger(ctx, slog.New(slog.NewTextHandler(os.Stdout, &slog.HandlerOptions{Level: level})))
+	log := Logger(ctx)
+
 	// Docker client should already be initialized and version negotiated in New()
 	if a.docker == nil {
 		return fmt.Errorf("docker client not initialized")
@@ -322,7 +339,7 @@ func (a *Agent) ExecTasks(ctx context.Context, jsonFilePath string) error {
 			return fmt.Errorf("error executing task %d: %w", taskRes.ID, err)
 		}
 
-		slog.Info("Task %d (Type: %d) executed successfully. Result: %+v", "task_id", taskRes.ID, "task_type", taskRes.TaskType, "result", result)
+		log.Info("task executed successfully", "task_id", taskRes.ID, "task_type", taskRes.TaskType, "result", result)
 	}
 
 	return nil

internal/app/ptah-agent/service.go

@@ -140,25 +140,34 @@ func (e *taskExecutor) prepareServicePayload(ctx context.Context, servicePayload
 		secret.SecretID = foundSecret.ID
 	}
 
-	if servicePayload.ReleaseCommand.Command != "" {
-		image, _, err := e.docker.ImageInspectWithRaw(ctx, spec.TaskTemplate.ContainerSpec.Image)
-		if err != nil {
-			return nil, errors.Wrapf(err, "get image %s", spec.TaskTemplate.ContainerSpec.Image)
-		}
+	image, _, err := e.docker.ImageInspectWithRaw(ctx, spec.TaskTemplate.ContainerSpec.Image)
+	if err != nil {
+		return nil, errors.Wrapf(err, "get image %s", spec.TaskTemplate.ContainerSpec.Image)
+	}
 
-		entrypoint := strings.Join(image.Config.Entrypoint, " ")
-		command := strings.Join(image.Config.Cmd, " ")
+	// FIXME: original entrypoint overrides custom command if both (release cmd & cmd) are set
+	entrypoint := strings.Join(image.Config.Entrypoint, " ")
+	command := strings.Join(image.Config.Cmd, " ")
 
-		originalEntrypoint := entrypoint + " " + command
+	spec.TaskTemplate.ContainerSpec.Env = append(spec.TaskTemplate.ContainerSpec.Env, fmt.Sprintf("ENTRYPOINT=%s %s", entrypoint, command))
 
+	if spec.TaskTemplate.ContainerSpec.Command != nil {
+		entrypoint = strings.Join(spec.TaskTemplate.ContainerSpec.Command, " ")
+	}
+
+	if spec.TaskTemplate.ContainerSpec.Args != nil {
+		command = strings.Join(spec.TaskTemplate.ContainerSpec.Args, " ")
+	}
+
+	if servicePayload.ReleaseCommand.Command != "" {
 		script := []string{
 			"#!/bin/sh",
 			"set -e",
 			"echo 'Starting release command'",
 			servicePayload.ReleaseCommand.Command,
 			"echo 'Release command finished'",
 			"echo 'Starting original entrypoint'",
-			originalEntrypoint,
+			fmt.Sprintf("exec %s %s", entrypoint, command),
 		}
 
 		config := swarm.ConfigSpec{

internal/app/ptah-agent/service_monitor.go

@@ -53,7 +53,7 @@ func (e *taskExecutor) monitorDaemonServiceLaunch(ctx context.Context, service *
 	defer ticker.Stop()
 
 	// TODO: make timeout configurable
-	timeout := time.After(time.Duration(10) * time.Minute)
+	timeout := time.After(time.Duration(1) * time.Minute)
 
 	successfullChecks := 0
 
@@ -106,6 +106,9 @@ func (e *taskExecutor) monitorDaemonServiceLaunch(ctx context.Context, service *
 				log.Debug("service inspected", "state", service.UpdateStatus.State)
 
 				switch service.UpdateStatus.State {
+				case swarm.UpdateStatePaused:
+					// FIXME: rollback the service?
+					return errors.Errorf("service update paused: %s", service.UpdateStatus.Message)
 				case swarm.UpdateStateCompleted:
 					return nil
 				case swarm.UpdateStateRollbackCompleted: