feat(deploy): print last logs from new container when fails to become healthy

cmd/uncloud/deploy.go

@@ -9,6 +9,7 @@ import (
 	"charm.land/lipgloss/v2"
 	composecli "github.com/compose-spec/compose-go/v2/cli"
 	"github.com/docker/compose/v2/pkg/progress"
+	"github.com/docker/docker/pkg/stringid"
 	"github.com/psviderski/uncloud/internal/cli"
 	"github.com/psviderski/uncloud/internal/cli/logs"
 	"github.com/psviderski/uncloud/internal/cli/tui"
@@ -20,10 +21,6 @@ import (
 	"github.com/spf13/cobra"
 )
 
-// failedContainerLogsTail is the number of recent log lines to print from a failed container to give the user immediate
-// context without requiring a follow-up 'uc logs' invocation.
-const failedContainerLogsTail = 10
-
 type deployOptions struct {
 	cli.BuildServicesOptions
 
@@ -238,34 +235,46 @@ func runDeploy(ctx context.Context, uncli *cli.CLI, opts deployOptions) error {
 	}, uncli.ProgressOut(), title)
 	if err != nil {
 		fmt.Println()
+
+		tail := failedContainerLogsTail()
 		if hookErr, ok := errors.AsType[*operation.PreDeployHookError](err); ok {
-			printPreDeployHookLogs(ctx, clusterClient, hookErr)
+			printFailedContainerLogs(ctx, clusterClient,
+				hookErr.ServiceName, hookErr.ContainerID, hookErr.MachineName, tail,
+				fmt.Sprintf("Last %d log lines from failed pre-deploy hook:", tail))
+			fmt.Println()
+		} else if startErr, ok := errors.AsType[*operation.ContainerHealthError](err); ok {
+			printFailedContainerLogs(ctx, clusterClient,
+				startErr.ServiceName, startErr.ContainerID, startErr.MachineName, tail,
+				fmt.Sprintf("Last %d log lines from failed container:", tail))
 			fmt.Println()
 		}
+
 		return err
 	}
 	return nil
 }
 
-// printPreDeployHookLogs fetches the last log lines from the failed pre-deploy hook container and prints them using
-// the standard log formatter.
-func printPreDeployHookLogs(ctx context.Context, cli *client.Client, hookErr *operation.PreDeployHookError) {
-	_, ch, err := cli.ServiceLogs(ctx, hookErr.ServiceName, api.ServiceLogsOptions{
-		Containers: []string{hookErr.ContainerID},
-		Machines:   []string{hookErr.MachineName},
-		Tail:       failedContainerLogsTail,
+// printFailedContainerLogs fetches the last tail log lines from a container that failed during deployment and prints
+// them using the standard log formatter under the provided header.
+func printFailedContainerLogs(
+	ctx context.Context, cli *client.Client, serviceName, containerID, machineName string, tail int, header string,
+) {
+	_, ch, err := cli.ServiceLogs(ctx, serviceName, api.ServiceLogsOptions{
+		Containers: []string{containerID},
+		Machines:   []string{machineName},
+		Tail:       tail,
 	})
 	if err != nil {
-		fmt.Fprintf(os.Stderr, "Failed to fetch pre-deploy hook logs: %v\n", err)
-		fmt.Fprintf(os.Stderr, "You can try manually with: uc logs %s\n", hookErr.ServiceName)
+		shortCtrID := stringid.TruncateID(containerID)
+		fmt.Fprintf(os.Stderr, "Failed to fetch container logs '%s/%s': %v\n", serviceName, shortCtrID, err)
+		fmt.Fprintf(os.Stderr, "You can try manually with: uc logs %s/%s\n", serviceName, shortCtrID)
 		return
 	}
 
-	header := fmt.Sprintf("Last %d log lines from failed pre-deploy hook:", failedContainerLogsTail)
 	fmt.Println(tui.BoldRed.Render(header))
 
 	logsEmpty := true
-	formatter := logs.NewFormatter([]string{hookErr.MachineName}, []string{hookErr.ServiceName}, false)
+	formatter := logs.NewFormatter([]string{machineName}, []string{serviceName}, false)
 	for entry := range ch {
 		logsEmpty = false
 		formatter.PrintEntry(entry)
@@ -275,3 +284,19 @@ func printPreDeployHookLogs(ctx context.Context, cli *client.Client, hookErr *op
 		fmt.Println("<no logs available>")
 	}
 }
+
+// defaultFailedContainerLogsTail is the default number of recent log lines to print from a failed container to give
+// the user immediate context without requiring a follow-up 'uc logs' invocation.
+// Overridable via UNCLOUD_FAILED_CONTAINER_LOGS_TAIL.
+const defaultFailedContainerLogsTail = 10
+
+// failedContainerLogsTail returns the number of log lines to fetch from a failed container, honouring the
+// UNCLOUD_FAILED_CONTAINER_LOGS_TAIL environment variable override when set and valid.
+func failedContainerLogsTail() int {
+	if v := os.Getenv("UNCLOUD_FAILED_CONTAINER_LOGS_TAIL"); v != "" {
+		if tail, err := logs.Tail(v); err == nil && (tail == -1 || tail > 0) {
+			return tail
+		}
+	}
+	return defaultFailedContainerLogsTail
+}

pkg/client/deploy/operation/container.go

@@ -13,6 +13,18 @@ import (
 	"github.com/psviderski/uncloud/pkg/api"
 )
 
+// ContainerHealthError indicates that a service container failed to become healthy during deployment.
+type ContainerHealthError struct {
+	error
+	ServiceName string
+	ContainerID string
+	MachineName string
+}
+
+func (e *ContainerHealthError) Unwrap() error {
+	return e.error
+}
+
 // RunContainerOperation creates and starts a new container on a specific machine.
 type RunContainerOperation struct {
 	ServiceID string
@@ -43,8 +55,13 @@ func (o *RunContainerOperation) Execute(ctx context.Context, cli Client) error {
 
 	opts := api.WaitContainerHealthyOptions{MonitorPeriod: o.Spec.UpdateConfig.MonitorPeriod}
 	if err = cli.WaitContainerHealthy(ctx, o.ServiceID, resp.ID, opts); err != nil {
-		return fmt.Errorf("container '%s/%s' failed to become healthy: %w",
-			o.Spec.Name, stringid.TruncateID(resp.ID), err)
+		return &ContainerHealthError{
+			error: fmt.Errorf("container '%s/%s' failed to become healthy: %w",
+				o.Spec.Name, stringid.TruncateID(resp.ID), err),
+			ServiceName: o.Spec.Name,
+			ContainerID: resp.ID,
+			MachineName: o.MachineName,
+		}
 	}
 
 	return nil
@@ -187,7 +204,6 @@ func (o *ReplaceContainerOperation) Execute(ctx context.Context, cli Client) err
 		if err = cli.WaitContainerHealthy(newCtx, o.ServiceID, resp.ID, opts); err != nil {
 			// New container failed to become healthy. Stop it and roll back to the previous container.
 			// Don't remove the new stopped container to allow users to inspect logs and state.
-			// TODO: collect logs from the new container and include in the error message to speed up debugging.
 
 			// Use context without progress to not overwrite the container Unhealthy status with Stopped.
 			ctxWithoutProgress := progress.WithContextWriter(ctx, nil)
@@ -196,21 +212,28 @@ func (o *ReplaceContainerOperation) Execute(ctx context.Context, cli Client) err
 			newCtr := fmt.Sprintf("%s/%s", o.Spec.Name, stringid.TruncateID(resp.ID))
 			healthErr := fmt.Errorf(
 				"new container '%s' failed to become healthy: %w. "+
-					"It's stopped and available for inspection. Fetch logs with 'uc logs %s'",
-				newCtr, err, o.Spec.Name,
+					"It's stopped and available for inspection. View logs with 'uc logs %s'",
+				newCtr, err, newCtr,
 			)
 
+			finalErr := healthErr
 			if stopFirst && wasRunning {
 				// Restart the old container only if it was running before we stopped it.
 				oldCtr := fmt.Sprintf("%s/%s", o.OldContainer.ServiceSpec.Name, o.OldContainer.ShortID())
 				if rollbackErr := cli.StartContainer(ctx, o.ServiceID, o.OldContainer.ID); rollbackErr != nil {
-					return fmt.Errorf("%w. Rolled back to old container '%s' but failed to restart it: %w",
+					finalErr = fmt.Errorf("%w. Rolled back to old container '%s' but failed to restart it: %w",
 						healthErr, oldCtr, rollbackErr)
+				} else {
+					finalErr = fmt.Errorf("%w. Rolled back to old container '%s'", healthErr, oldCtr)
 				}
-				return fmt.Errorf("%w. Rolled back to old container '%s'", healthErr, oldCtr)
 			}
 
-			return healthErr
+			return &ContainerHealthError{
+				error:       finalErr,
+				ServiceName: o.Spec.Name,
+				ContainerID: resp.ID,
+				MachineName: o.MachineName,
+			}
 		}
 	}
 

pkg/client/deploy/operation/predeploy.go

@@ -24,6 +24,10 @@ type PreDeployHookError struct {
 	MachineName string
 }
 
+func (e *PreDeployHookError) Unwrap() error {
+	return e.error
+}
+
 // DefaultPreDeployTimeout is the maximum duration to wait for a pre-deploy hook container to complete.
 const DefaultPreDeployTimeout = 5 * time.Minute