CLI: Return non-zero exit code when deployment fails in nomad run

.changelog/11550.txt

@@ -0,0 +1,3 @@
+```release-note:improvement
+cli: Return non-zero exit code from monitor if deployment fails
+```

command/deployment_status.go

@@ -181,11 +181,11 @@ func (c *DeploymentStatusCommand) Run(args []string) int {
 	return 0
 }
 
-func (c *DeploymentStatusCommand) monitor(client *api.Client, deployID string, index uint64, verbose bool) {
+func (c *DeploymentStatusCommand) monitor(client *api.Client, deployID string, index uint64, verbose bool) (status string, err error) {
 	if isStdoutTerminal() {
-		c.ttyMonitor(client, deployID, index, verbose)
+		return c.ttyMonitor(client, deployID, index, verbose)
 	} else {
-		c.defaultMonitor(client, deployID, index, verbose)
+		return c.defaultMonitor(client, deployID, index, verbose)
 	}
 }
 
@@ -208,7 +208,7 @@ func isStdoutTerminal() bool {
 // but only used for tty and non-Windows machines since glint doesn't work with
 // cmd/PowerShell and non-interactive interfaces
 // Margins are used to match the text alignment from job run
-func (c *DeploymentStatusCommand) ttyMonitor(client *api.Client, deployID string, index uint64, verbose bool) {
+func (c *DeploymentStatusCommand) ttyMonitor(client *api.Client, deployID string, index uint64, verbose bool) (status string, err error) {
 	var length int
 	if verbose {
 		length = fullId
@@ -242,7 +242,9 @@ func (c *DeploymentStatusCommand) ttyMonitor(client *api.Client, deployID string
 
 UPDATE:
 	for {
-		deploy, meta, err := client.Deployments().Info(deployID, &q)
+		var deploy *api.Deployment
+		var meta *api.QueryMeta
+		deploy, meta, err = client.Deployments().Info(deployID, &q)
 		if err != nil {
 			d.Append(glint.Layout(glint.Style(
 				glint.Text(fmt.Sprintf("%s: Error fetching deployment", formatTime(time.Now()))),
@@ -252,7 +254,7 @@ UPDATE:
 			return
 		}
 
-		status := deploy.Status
+		status = deploy.Status
 		statusComponent = glint.Layout(
 			glint.Text(""),
 			glint.Text(formatTime(time.Now())),
@@ -309,7 +311,8 @@ UPDATE:
 
 				// Wait for rollback to launch
 				time.Sleep(1 * time.Second)
-				rollback, _, err := client.Jobs().LatestDeployment(deploy.JobID, nil)
+				var rollback *api.Deployment
+				rollback, _, err = client.Jobs().LatestDeployment(deploy.JobID, nil)
 
 				if err != nil {
 					d.Append(glint.Layout(glint.Style(
@@ -342,7 +345,7 @@ UPDATE:
 				glint.Text(fmt.Sprintf("✓ Deployment %q %s", limit(deployID, length), status)),
 			).Row().MarginLeft(2)
 			break UPDATE
-		case structs.DeploymentStatusCancelled, structs.DeploymentStatusDescriptionBlocked:
+		case structs.DeploymentStatusCancelled, structs.DeploymentStatusBlocked:
 			endSpinner = glint.Layout(
 				glint.Text(fmt.Sprintf("! Deployment %q %s", limit(deployID, length), status)),
 			).Row().MarginLeft(2)
@@ -355,10 +358,11 @@ UPDATE:
 	// Render one final time with completion message
 	d.Set(endSpinner, statusComponent, glint.Text(""))
 	d.RenderFrame()
+	return
 }
 
 // Used for Windows and non-tty
-func (c *DeploymentStatusCommand) defaultMonitor(client *api.Client, deployID string, index uint64, verbose bool) {
+func (c *DeploymentStatusCommand) defaultMonitor(client *api.Client, deployID string, index uint64, verbose bool) (status string, err error) {
 	writer := uilive.New()
 	writer.Start()
 	defer writer.Stop()
@@ -377,13 +381,15 @@ func (c *DeploymentStatusCommand) defaultMonitor(client *api.Client, deployID st
 	}
 
 	for {
-		deploy, meta, err := client.Deployments().Info(deployID, &q)
+		var deploy *api.Deployment
+		var meta *api.QueryMeta
+		deploy, meta, err = client.Deployments().Info(deployID, &q)
 		if err != nil {
 			c.Ui.Error(c.Colorize().Color(fmt.Sprintf("%s: Error fetching deployment", formatTime(time.Now()))))
 			return
 		}
 
-		status := deploy.Status
+		status = deploy.Status
 		info := formatTime(time.Now())
 		info += fmt.Sprintf("\n%s", formatDeployment(client, deploy, length))
 
@@ -413,7 +419,8 @@ func (c *DeploymentStatusCommand) defaultMonitor(client *api.Client, deployID st
 			if hasAutoRevert(deploy) {
 				// Wait for rollback to launch
 				time.Sleep(1 * time.Second)
-				rollback, _, err := client.Jobs().LatestDeployment(deploy.JobID, nil)
+				var rollback *api.Deployment
+				rollback, _, err = client.Jobs().LatestDeployment(deploy.JobID, nil)
 
 				// Separate rollback monitoring from failed deployment
 				// Needs to be after time.Sleep or it messes up the formatting
@@ -436,7 +443,7 @@ func (c *DeploymentStatusCommand) defaultMonitor(client *api.Client, deployID st
 			}
 			return
 
-		case structs.DeploymentStatusSuccessful, structs.DeploymentStatusCancelled, structs.DeploymentStatusDescriptionBlocked:
+		case structs.DeploymentStatusSuccessful, structs.DeploymentStatusCancelled, structs.DeploymentStatusBlocked:
 			return
 		default:
 			q.WaitIndex = meta.LastIndex

command/monitor.go

@@ -300,7 +300,10 @@ func (m *monitor) monitor(evalID string) int {
 		meta := new(Meta)
 		meta.Ui = m.ui
 		cmd := &DeploymentStatusCommand{Meta: *meta}
-		cmd.monitor(m.client, dID, 0, verbose)
+		status, err := cmd.monitor(m.client, dID, 0, verbose)
+		if err != nil || status != structs.DeploymentStatusSuccessful {
+			return 1
+		}
 	}
 
 	// Treat scheduling failures specially using a dedicated exit code.

website/content/docs/commands/job/run.mdx

@@ -34,7 +34,8 @@ exit after scheduling and deployment have finished or failed.
 On successful job submission and scheduling, exit code 0 will be returned. If
 there are job placement issues encountered (unsatisfiable constraints, resource
 exhaustion, etc), then the exit code will be 2. Any other errors, including
-client connection issues or internal errors, are indicated by exit code 1.
+deployment failures, client connection issues, or internal errors, are indicated
+by exit code 1.
 
 If the job has specified the region, the `-region` flag and `$NOMAD_REGION`
 environment variable are overridden and the job's region is used.