feat(deploy): per-service deploy queue, cancellable deployments, no-cache builds

- internal/deployqueue: serialized FIFO per service — concurrent deploys
  of one service no longer race (the container-name-conflict failure
  mode); different services still run in parallel.
- Deploys and rollbacks enqueue through the queue; a busy service
  returns status 'queued' and promotes when the active job finishes.
- POST /deployments/:id/cancel — queued jobs drop immediately
  (200 cancelled); active jobs get ctx cancelled so build/reconcile
  abort (202 cancelling, settles 'cancelled'); finished deployments
  return 409 NOT_CANCELLABLE.
- Redeploy and variables save now route through the same queue so they
  can't overlap a build; PUT /services/:id/variables accepts
  redeploy:true to apply env changes to running containers.
- no_cache on deployment create plumbs BuildRequest -> BuildOptions for
  nixpacks, dockerfile and railpack builders.
- CLI: deployments cancel, deploy --no-cache, variables set/unset
  --redeploy. MCP: containr_deployments_cancel; deploy/variables_set
  document no_cache/redeploy. UI: Cancel button + queued/cancelled
  badges on the deployments table. OpenAPI + types regenerated.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

AGENTS.md

@@ -28,6 +28,7 @@ Dev stack: Postgres on `:15432` (`containr-postgres-1`), Dragonfly on `:16379` (
 ## Conventions that matter
 
 - New endpoints go in the right route group; mutations on `authed`/`admin` get idempotency for free.
+- Deploy work goes through `deploy_queue` (`internal/deployqueue`) — per-service FIFO; enqueue jobs instead of `go`-ing deployment/reconcile goroutines so they never race a build.
 - New list endpoints use `pageWindow`/`paginationMeta` (`internal/api/pagination.go`) — cursor + page, `next_cursor` in the envelope.
 - New errors use stable codes via `respondError(c, status, code, msg)` in `internal/api/errors.go`.
 - New CLI command → also register an MCP tool in `internal/mcpserver/server.go` and document in OpenAPI. All three surfaces ship together.

app/backend/internal/api/deployments.go

@@ -3,7 +3,9 @@ package api
 import (
 	"containr/internal/database"
 	"containr/internal/deployment"
+	"containr/internal/deployqueue"
 	"context"
+	"errors"
 	"fmt"
 	"net/http"
 	"strings"
@@ -34,6 +36,7 @@ type CreateDeploymentRequest struct {
 	Branch     string            `json:"branch"`
 	Trigger    string            `json:"trigger"`
 	EnvVars    map[string]string `json:"env_vars"`
+	NoCache    bool              `json:"no_cache"`
 }
 
 type DeploymentResponse struct {
@@ -309,7 +312,23 @@ func handleCreateDeployment(c *gin.Context) {
 		}
 
 		engineInstance := engine.(*deployment.DeploymentEngine)
-		go runDeploymentAndSync(context.Background(), db.(*database.DB), engineInstance, &d, service, req, userID.(string))
+		pos := getDeployQueue(c).Enqueue(serviceID, deployqueue.Job{
+			DeploymentID: d.ID,
+			Run: func(jctx context.Context) {
+				runDeploymentAndSync(jctx, db.(*database.DB), engineInstance, &d, service, req, userID.(string))
+			},
+		})
+		if pos > 0 {
+			d.Status = "queued"
+			_, _ = db.(*database.DB).Exec(
+				`UPDATE deployments SET status = 'queued', updated_at = $1 WHERE id = $2`,
+				time.Now(), d.ID,
+			)
+			_, _ = db.(*database.DB).Exec(
+				`UPDATE services SET status = 'queued', updated_at = $1 WHERE id = $2`,
+				time.Now(), serviceID,
+			)
+		}
 	}
 
 	c.JSON(http.StatusCreated, DeploymentResponse{
@@ -427,6 +446,7 @@ func runDeploymentAndSyncWithImage(
 			SourcePath: sourcePath,
 			Branch:     req.Branch,
 			Commit:     req.CommitHash,
+			NoCache:    req.NoCache,
 		}
 	}
 
@@ -454,16 +474,19 @@ func runDeploymentAndSyncWithImage(
 		select {
 		case <-ctx.Done():
 			failedAt := time.Now()
-			timeoutErr := "Deployment timed out before completion"
+			finalStatus, finalErr := "failed", "Deployment timed out before completion"
+			if errors.Is(ctx.Err(), context.Canceled) {
+				finalStatus, finalErr = "cancelled", "Deployment cancelled"
+			}
 			_, _ = db.Exec(
 				`UPDATE deployments
-				 SET status = 'failed', error = $1, completed_at = $2, updated_at = $2
-				 WHERE id = $3`,
-				timeoutErr, failedAt, dbDeployment.ID,
+				 SET status = $1, error = $2, completed_at = $3, updated_at = $3
+				 WHERE id = $4`,
+				finalStatus, finalErr, failedAt, dbDeployment.ID,
 			)
 			_, _ = db.Exec(
-				`UPDATE services SET status = 'failed', updated_at = $1 WHERE id = $2`,
-				failedAt, service.ID,
+				`UPDATE services SET status = $1, updated_at = $2 WHERE id = $3`,
+				finalStatus, failedAt, service.ID,
 			)
 			return
 		case <-syncTicker.C:
@@ -530,13 +553,86 @@ func runDeploymentAndSyncWithImage(
 }
 
 func mapEngineStatusToDBStatus(status string) string {
-	switch status {
-	case "running":
+	if status == "running" {
 		return "deployed"
-	case "cancelled":
-		return "failed"
-	default:
+	}
 	return status
+}
+
+// getDeployQueue returns the per-service deploy queue. SetupRoutes always
+// sets it; the fallback keeps handler tests that skip SetupRoutes working.
+func getDeployQueue(c *gin.Context) *deployqueue.Queue {
+	if v, exists := c.Get("deploy_queue"); exists {
+		if q, ok := v.(*deployqueue.Queue); ok && q != nil {
+			return q
+		}
+	}
+	return deployqueue.New()
+}
+
+// handleCancelDeployment cancels a queued or in-flight deployment. Queued
+// jobs are dropped and marked cancelled immediately; active jobs get their
+// context cancelled so the build/reconcile aborts and the sync loop writes
+// the terminal 'cancelled' status.
+func handleCancelDeployment(c *gin.Context) {
+	db, exists := c.Get("db")
+	if !exists {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": "Database connection not available", "code": "DEPENDENCY_UNAVAILABLE"})
+		return
+	}
+
+	deploymentID, err := uuid.Parse(c.Param("id"))
+	if err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "Invalid deployment ID", "code": "VALIDATION"})
+		return
+	}
+	userID, exists := c.Get("user_id")
+	if !exists {
+		c.JSON(http.StatusUnauthorized, gin.H{"error": "User not authenticated", "code": "UNAUTHENTICATED"})
+		return
+	}
+
+	var status, owner string
+	var serviceID uuid.UUID
+	err = db.(*database.DB).QueryRow(
+		`SELECT d.status, d.service_id, p.owner_id
+		 FROM deployments d
+		 JOIN services s ON d.service_id = s.id
+		 JOIN projects p ON s.project_id = p.id
+		 WHERE d.id = $1`,
+		deploymentID,
+	).Scan(&status, &serviceID, &owner)
+	if err != nil {
+		c.JSON(http.StatusNotFound, gin.H{"error": "Deployment not found", "code": "NOT_FOUND"})
+		return
+	}
+	if owner != userID.(string) && !contextIsAdmin(c) {
+		c.JSON(http.StatusForbidden, gin.H{"error": "Access denied", "code": "FORBIDDEN"})
+		return
+	}
+
+	switch status {
+	case "deployed", "failed", "cancelled":
+		c.JSON(http.StatusConflict, gin.H{"error": "Deployment already finished", "code": "NOT_CANCELLABLE"})
+		return
+	}
+
+	switch getDeployQueue(c).Cancel(serviceID, deploymentID) {
+	case deployqueue.WasQueued:
+		now := time.Now()
+		_, _ = db.(*database.DB).Exec(
+			`UPDATE deployments SET status = 'cancelled', error = 'Deployment cancelled', completed_at = $1, updated_at = $1 WHERE id = $2`,
+			now, deploymentID,
+		)
+		_, _ = db.(*database.DB).Exec(
+			`UPDATE services SET status = 'cancelled', updated_at = $1 WHERE id = $2`,
+			now, serviceID,
+		)
+		c.JSON(http.StatusOK, gin.H{"status": "cancelled"})
+	case deployqueue.WasActive:
+		c.JSON(http.StatusAccepted, gin.H{"status": "cancelling"})
+	default:
+		c.JSON(http.StatusConflict, gin.H{"error": "Deployment is not running on this node", "code": "NOT_CANCELLABLE"})
 	}
 }
 
@@ -735,7 +831,12 @@ func handleRollbackDeployment(c *gin.Context) {
 			Trigger: "rollback",
 		}
 		// Reuse the stored image directly — no rebuild.
-		go runDeploymentAndSyncWithImage(context.Background(), db.(*database.DB), engine, &rollback, service, rollbackReq, userID.(string), targetImage)
+		getDeployQueue(c).Enqueue(serviceID, deployqueue.Job{
+			DeploymentID: rollback.ID,
+			Run: func(jctx context.Context) {
+				runDeploymentAndSyncWithImage(jctx, db.(*database.DB), engine, &rollback, service, rollbackReq, userID.(string), targetImage)
+			},
+		})
 	}
 
 	c.JSON(http.StatusCreated, gin.H{

app/backend/internal/api/routes.go

@@ -10,6 +10,7 @@ import (
 	"containr/internal/config"
 	"containr/internal/database"
 	"containr/internal/deployment"
+	"containr/internal/deployqueue"
 	"containr/internal/docker"
 	"containr/internal/ha"
 	"containr/internal/metrics"
@@ -98,6 +99,10 @@ func SetupRoutes(router *gin.Engine, db *database.DB, redis *database.Redis, cfg
 	// Initialize database handler
 	databaseHandler := NewDatabaseHandler(db.DB, dockerClient)
 
+	// Per-service serialized deploy queue — one active deployment per
+	// service, the rest wait as 'queued'.
+	deployQueue := deployqueue.New()
+
 	// Initialize security handler
 	securityHandler := NewSecurityHandler(db, cfg.JWTSecret)
 
@@ -115,6 +120,7 @@ func SetupRoutes(router *gin.Engine, db *database.DB, redis *database.Redis, cfg
 		if deploymentEngine != nil {
 			c.Set("deployment_engine", deploymentEngine)
 		}
+		c.Set("deploy_queue", deployQueue)
 		c.Set("scheduler", scheduler)
 		c.Set("metrics_collector", metricsCollector)
 		c.Set("auto_scaler", autoScaler)
@@ -289,6 +295,7 @@ func SetupRoutes(router *gin.Engine, db *database.DB, redis *database.Redis, cfg
 			authed.POST("/services/:id/redeploy", handleServiceRedeploy)
 			authed.POST("/services/:id/deployments", handleCreateDeployment)
 			authed.POST("/deployments/:id/rollback", handleRollbackDeployment)
+			authed.POST("/deployments/:id/cancel", handleCancelDeployment)
 
 			// Environment variables — secrets, owner/admin only.
 			authed.GET("/services/:id/variables", handleGetVariables)

app/backend/internal/api/runtime.go

@@ -11,6 +11,7 @@ import (
 
 	"containr/internal/database"
 	"containr/internal/deployment"
+	"containr/internal/deployqueue"
 
 	"github.com/gin-gonic/gin"
 	"github.com/google/uuid"
@@ -381,6 +382,7 @@ func handleServiceRestart(c *gin.Context) {
 
 // handleServiceRedeploy re-applies the current spec: pulls the image for
 // image-sourced services and reconciles containers (env, ports, replicas).
+// Runs through the per-service deploy queue so it never overlaps a build.
 func handleServiceRedeploy(c *gin.Context) {
 	engine, db, ok := getRuntimeEngine(c)
 	if !ok {
@@ -406,7 +408,78 @@ func handleServiceRedeploy(c *gin.Context) {
 		}
 	}
 	service.Image = image
-	reconcileNow(c, engine, db, service)
+
+	finished := make(chan struct{})
+	var recErr error
+	var state *deployment.RuntimeState
+	getDeployQueue(c).Enqueue(service.ID, deployqueue.Job{Run: func(jctx context.Context) {
+		defer close(finished)
+		state, recErr = reconcileServiceJob(jctx, db, engine, service)
+	}})
+
+	select {
+	case <-finished:
+		if recErr != nil {
+			c.JSON(http.StatusBadGateway, gin.H{"error": recErr.Error()})
+			return
+		}
+		c.JSON(http.StatusOK, gin.H{"runtime": state})
+	case <-c.Request.Context().Done():
+		c.JSON(http.StatusAccepted, gin.H{"status": "queued"})
+	}
+}
+
+// reconcileServiceJob is the shared queue-job body for redeploy paths:
+// build the spec from current DB state, reconcile containers, persist the
+// resulting status and published port.
+func reconcileServiceJob(ctx context.Context, db *database.DB, engine *deployment.DeploymentEngine, service Service) (*deployment.RuntimeState, error) {
+	spec, err := serviceRuntimeSpec(db, service)
+	if err != nil {
+		return nil, err
+	}
+	state, err := engine.ReconcileService(ctx, spec)
+	if err != nil {
+		return nil, err
+	}
+	if state != nil {
+		persistPublishedPort(db, service.ID, state.Ports)
+		status := "stopped"
+		if state.Status != "" {
+			status = state.Status
+		}
+		_, _ = db.Exec(`UPDATE services SET status = $1, updated_at = $2 WHERE id = $3`, status, time.Now(), service.ID)
+	}
+	return state, nil
+}
+
+// enqueueServiceRedeploy resolves a deployable image (service.Image, else
+// the last successful deployment's tag) and queues a reconcile. Returns
+// false when no image is available or the engine is missing.
+func enqueueServiceRedeploy(c *gin.Context, db *database.DB, service Service) bool {
+	engineValue, exists := c.Get("deployment_engine")
+	if !exists || engineValue == nil {
+		return false
+	}
+	engine, ok := engineValue.(*deployment.DeploymentEngine)
+	if !ok {
+		return false
+	}
+	image := service.Image
+	if image == "" {
+		if err := db.QueryRow(
+			`SELECT image_name || ':' || image_tag FROM deployments
+			 WHERE service_id = $1 AND status = 'deployed' AND image_name <> ''
+			 ORDER BY created_at DESC LIMIT 1`,
+			service.ID,
+		).Scan(&image); err != nil {
+			return false
+		}
+	}
+	service.Image = image
+	getDeployQueue(c).Enqueue(service.ID, deployqueue.Job{Run: func(jctx context.Context) {
+		_, _ = reconcileServiceJob(jctx, db, engine, service)
+	}})
+	return true
 }
 
 // persistPublishedPort records the host port Docker actually bound so later

app/backend/internal/api/variables.go

@@ -25,6 +25,7 @@ type EnvironmentVariable struct {
 
 type UpdateVariablesRequest struct {
 	Variables []VariableInput `json:"variables" binding:"required"`
+	Redeploy  bool            `json:"redeploy"`
 }
 
 type VariableInput struct {
@@ -204,5 +205,22 @@ func handleUpdateVariables(c *gin.Context) {
 		variables = append(variables, v)
 	}
 
-	c.JSON(http.StatusOK, gin.H{"variables": variables, "message": "Environment variables updated successfully"})
+	resp := gin.H{"variables": variables, "message": "Environment variables updated successfully"}
+	if req.Redeploy {
+		// Load the service spec inputs and queue a reconcile behind any
+		// in-flight deployment — new env lands in the recreated containers.
+		var svc Service
+		svcErr := db.(*database.DB).QueryRow(
+			`SELECT id, project_id, name, COALESCE(image, ''), COALESCE(command, ''),
+			        COALESCE(replicas, 1), COALESCE(port, 0), COALESCE(domain, ''),
+			        COALESCE(healthcheck_path, ''), COALESCE(restart_policy, 'unless-stopped'),
+			        COALESCE(cpu, ''), COALESCE(memory, '')
+			 FROM services WHERE id = $1`,
+			serviceID,
+		).Scan(&svc.ID, &svc.ProjectID, &svc.Name, &svc.Image, &svc.Command,
+			&svc.Replicas, &svc.Port, &svc.Domain, &svc.HealthCheckPath,
+			&svc.RestartPolicy, &svc.CPU, &svc.Memory)
+		resp["redeploy_queued"] = svcErr == nil && enqueueServiceRedeploy(c, db.(*database.DB), svc)
+	}
+	c.JSON(http.StatusOK, resp)
 }

app/backend/internal/build/dockerfile.go

@@ -222,6 +222,7 @@ func (d *DockerfileBuilder) Build(ctx context.Context, req *types.BuildRequest)
 		Dockerfile: "Dockerfile",
 		Tags:       []string{imageName},
 		BuildArgs:  buildArgs,
+		NoCache:    req.NoCache,
 		Remove:     true,
 	}
 

app/backend/internal/build/nixpacks.go

@@ -498,6 +498,7 @@ func (n *NixpacksBuilder) Build(ctx context.Context, req *types.BuildRequest) (*
 		BuildArgs: map[string]*string{
 			"BUILDKIT_INLINE_CACHE": strPtr("1"),
 		},
+		NoCache: req.NoCache,
 		Remove:  true,
 	}
 

app/backend/internal/build/railpack.go

@@ -156,6 +156,7 @@ func (rb *RailpackBuilder) Build(ctx context.Context, req *types.BuildRequest) (
 		Tags:       []string{imageName},
 		BuildArgs:  buildArgs,
 		Labels:     req.Labels,
+		NoCache:    req.NoCache,
 		Remove:     true,
 	}
 

app/backend/internal/cli/commands/deploy.go

@@ -17,6 +17,9 @@ var DeployCmd = &cobra.Command{
 			return err
 		}
 		body := map[string]interface{}{"trigger": "cli"}
+		if nc, _ := cmd.Flags().GetBool("no-cache"); nc {
+			body["no_cache"] = true
+		}
 		for _, f := range []string{"commit-hash", "branch"} {
 			if v, _ := cmd.Flags().GetString(f); v != "" {
 				key := "commit_hash"
@@ -106,6 +109,29 @@ var deploymentsLogsCmd = &cobra.Command{
 	RunE:  runDeploymentLogs,
 }
 
+var deploymentsCancelCmd = &cobra.Command{
+	Use:   "cancel <id>",
+	Short: "Cancel a queued or running deployment",
+	Args:  cobra.ExactArgs(1),
+	RunE: func(cmd *cobra.Command, args []string) error {
+		c, err := client()
+		if err != nil {
+			return err
+		}
+		data, err := c.Do("POST", "/deployments/"+args[0]+"/cancel", map[string]interface{}{})
+		if err != nil {
+			return err
+		}
+		if JSONMode() {
+			PrintRaw(data)
+			return nil
+		}
+		obj, _ := unwrapObject(data)
+		fmt.Printf("Deployment %s: %s\n", args[0], str(obj, "status"))
+		return nil
+	},
+}
+
 var deploymentsRollbackCmd = &cobra.Command{
 	Use:   "rollback <id>",
 	Short: "Roll back to a deployment",
@@ -131,5 +157,6 @@ var deploymentsRollbackCmd = &cobra.Command{
 func init() {
 	DeployCmd.Flags().String("commit-hash", "", "commit to deploy")
 	DeployCmd.Flags().String("branch", "", "branch to deploy")
-	DeploymentsCmd.AddCommand(deploymentsListCmd, deploymentsGetCmd, deploymentsLogsCmd, deploymentsRollbackCmd)
+	DeployCmd.Flags().Bool("no-cache", false, "build without Docker layer cache")
+	DeploymentsCmd.AddCommand(deploymentsListCmd, deploymentsGetCmd, deploymentsLogsCmd, deploymentsRollbackCmd, deploymentsCancelCmd)
 }

app/backend/internal/cli/commands/variables.go

@@ -49,7 +49,8 @@ var varsSetCmd = &cobra.Command{
 	Args:  cobra.MinimumNArgs(2),
 	RunE: func(cmd *cobra.Command, args []string) error {
 		secret, _ := cmd.Flags().GetBool("secret")
-		return mutateVars(args[0], func(vars []map[string]interface{}) []map[string]interface{} {
+		redeploy, _ := cmd.Flags().GetBool("redeploy")
+		return mutateVars(args[0], redeploy, func(vars []map[string]interface{}) []map[string]interface{} {
 			for _, pair := range args[1:] {
 				k, v, found := strings.Cut(pair, "=")
 				if !found || k == "" {
@@ -71,7 +72,8 @@ var varsUnsetCmd = &cobra.Command{
 		for _, k := range args[1:] {
 			drop[k] = true
 		}
-		return mutateVars(args[0], func(vars []map[string]interface{}) []map[string]interface{} {
+		redeploy, _ := cmd.Flags().GetBool("redeploy")
+		return mutateVars(args[0], redeploy, func(vars []map[string]interface{}) []map[string]interface{} {
 			out := vars[:0]
 			for _, v := range vars {
 				if !drop[str(v, "key")] {
@@ -91,7 +93,7 @@ func fetchVars(c *Client, serviceID string) ([]map[string]interface{}, error) {
 	return unwrapList(data, "variables")
 }
 
-func mutateVars(serviceID string, mutate func([]map[string]interface{}) []map[string]interface{}) error {
+func mutateVars(serviceID string, redeploy bool, mutate func([]map[string]interface{}) []map[string]interface{}) error {
 	c, err := client()
 	if err != nil {
 		return err
@@ -109,7 +111,11 @@ func mutateVars(serviceID string, mutate func([]map[string]interface{}) []map[st
 			"is_secret": v["is_secret"] == true,
 		})
 	}
-	data, err := c.Do("PUT", "/services/"+serviceID+"/variables", map[string]interface{}{"variables": payload})
+	body := map[string]interface{}{"variables": payload}
+	if redeploy {
+		body["redeploy"] = true
+	}
+	data, err := c.Do("PUT", "/services/"+serviceID+"/variables", body)
 	if err != nil {
 		return err
 	}
@@ -117,7 +123,12 @@ func mutateVars(serviceID string, mutate func([]map[string]interface{}) []map[st
 		PrintRaw(data)
 		return nil
 	}
+	obj, _ := unwrapObject(data)
+	if obj["redeploy_queued"] == true {
+		fmt.Printf("Variables updated on %s — redeploy queued\n", serviceID)
+	} else {
 		fmt.Printf("Variables updated on %s — redeploy to apply to running containers\n", serviceID)
+	}
 	return nil
 }
 
@@ -134,5 +145,7 @@ func upsertVar(vars []map[string]interface{}, key, value string, secret bool) []
 
 func init() {
 	varsSetCmd.Flags().Bool("secret", false, "mark the variable(s) as secret")
+	varsSetCmd.Flags().Bool("redeploy", false, "redeploy the service after saving")
+	varsUnsetCmd.Flags().Bool("redeploy", false, "redeploy the service after saving")
 	VariablesCmd.AddCommand(varsListCmd, varsSetCmd, varsUnsetCmd)
 }

app/backend/internal/deployment/engine.go

@@ -150,6 +150,7 @@ type BuildConfig struct {
 	Environment   map[string]string `json:"environment"`
 	Branch        string            `json:"branch"`
 	Commit        string            `json:"commit"`
+	NoCache       bool              `json:"no_cache"`
 }
 
 type TriggerConfig struct {
@@ -291,6 +292,7 @@ func (de *DeploymentEngine) buildImage(ctx context.Context, deployment *Deployme
 		TriggeredBy:   "deployment_engine",
 		Branch:        buildConfig.Branch,
 		Commit:        buildConfig.Commit,
+		NoCache:       buildConfig.NoCache,
 	}
 
 	response, err := de.buildManager.Build(ctx, buildReq)

app/backend/internal/deployqueue/queue.go

@@ -0,0 +1,146 @@
+package deployqueue
+
+import (
+	"context"
+	"sync"
+
+	"github.com/google/uuid"
+)
+
+// Job is a unit of serialized deployment work. DeploymentID is the
+// deployments row it mirrors (uuid.Nil for runtime ops with no row).
+// Run is invoked with a cancellable context — cancelled by Cancel for
+// active jobs, or the job is dropped before starting when queued.
+type Job struct {
+	DeploymentID uuid.UUID
+	Run          func(ctx context.Context)
+}
+
+// Queue serializes deploy work per service: one active job, the rest FIFO.
+// A build for service A never races a redeploy or second build of A, while
+// services B, C proceed in parallel. Not a global worker pool — per-service
+// serialization is what prevents container thrash.
+type Queue struct {
+	mu      sync.Mutex
+	states  map[uuid.UUID]*serviceState
+	cancels map[uuid.UUID]context.CancelFunc
+}
+
+type serviceState struct {
+	running bool
+	pending []Job
+}
+
+func New() *Queue {
+	return &Queue{
+		states:  make(map[uuid.UUID]*serviceState),
+		cancels: make(map[uuid.UUID]context.CancelFunc),
+	}
+}
+
+// Enqueue schedules a job. Returns the position: 0 means it starts
+// immediately, >0 means that many jobs run first.
+func (q *Queue) Enqueue(serviceID uuid.UUID, job Job) int {
+	q.mu.Lock()
+	st := q.states[serviceID]
+	if st == nil {
+		st = &serviceState{}
+		q.states[serviceID] = st
+	}
+	if st.running {
+		st.pending = append(st.pending, job)
+		pos := len(st.pending)
+		q.mu.Unlock()
+		return pos
+	}
+	st.running = true
+	q.mu.Unlock()
+
+	q.start(serviceID, job)
+	return 0
+}
+
+func (q *Queue) start(serviceID uuid.UUID, job Job) {
+	ctx, cancel := context.WithCancel(context.Background())
+	if job.DeploymentID != uuid.Nil {
+		q.mu.Lock()
+		q.cancels[job.DeploymentID] = cancel
+		q.mu.Unlock()
+	}
+
+	go func() {
+		defer func() {
+			if job.DeploymentID != uuid.Nil {
+				q.mu.Lock()
+				delete(q.cancels, job.DeploymentID)
+				q.mu.Unlock()
+			}
+			cancel()
+			q.promote(serviceID)
+		}()
+		job.Run(ctx)
+	}()
+}
+
+// promote marks the service free and starts the next queued job, if any.
+func (q *Queue) promote(serviceID uuid.UUID) {
+	q.mu.Lock()
+	st := q.states[serviceID]
+	if st == nil || len(st.pending) == 0 {
+		if st != nil {
+			st.running = false
+		}
+		q.mu.Unlock()
+		return
+	}
+	next := st.pending[0]
+	st.pending = st.pending[1:]
+	q.mu.Unlock()
+	q.start(serviceID, next)
+}
+
+// CancelState distinguishes what Cancel found.
+type CancelState int
+
+const (
+	NotFound  CancelState = iota
+	WasQueued             // removed before it started — caller marks DB cancelled
+	WasActive             // ctx cancelled — the running job marks DB itself
+)
+
+// Cancel cancels a job by its deployment row id.
+func (q *Queue) Cancel(serviceID, deploymentID uuid.UUID) CancelState {
+	q.mu.Lock()
+	if st := q.states[serviceID]; st != nil {
+		for i, j := range st.pending {
+			if j.DeploymentID == deploymentID {
+				st.pending = append(st.pending[:i], st.pending[i+1:]...)
+				q.mu.Unlock()
+				return WasQueued
+			}
+		}
+	}
+	cancel, ok := q.cancels[deploymentID]
+	q.mu.Unlock()
+	if ok {
+		cancel()
+		return WasActive
+	}
+	return NotFound
+}
+
+// Do runs fn through the queue and waits for completion — for synchronous
+// endpoints (redeploy) that still must not overlap an in-flight deployment.
+// ctx bounds the wait, not the job; on timeout the job still runs to
+// completion to avoid orphaned reconcile state.
+func (q *Queue) Do(ctx context.Context, serviceID uuid.UUID, fn func(ctx context.Context)) {
+	done := make(chan struct{})
+	q.Enqueue(serviceID, Job{Run: func(jctx context.Context) {
+		defer close(done)
+		fn(jctx)
+	}})
+	select {
+	case <-done:
+	case <-ctx.Done():
+	}
+}

app/backend/internal/deployqueue/queue_test.go

@@ -0,0 +1,142 @@
+package deployqueue
+
+import (
+	"context"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+)
+
+func TestSerializesPerService(t *testing.T) {
+	q := New()
+	svc := uuid.New()
+
+	var active, maxActive int32
+	finish := make(chan struct{})
+	for i := 0; i < 3; i++ {
+		q.Enqueue(svc, Job{Run: func(ctx context.Context) {
+			if n := atomic.AddInt32(&active, 1); n > atomic.LoadInt32(&maxActive) {
+				atomic.StoreInt32(&maxActive, n)
+			}
+			<-finish // hold until released
+			atomic.AddInt32(&active, -1)
+		}})
+	}
+
+	deadline := time.After(2 * time.Second)
+	for atomic.LoadInt32(&active) < 1 {
+		select {
+		case <-deadline:
+			t.Fatal("first job never started")
+		default:
+			time.Sleep(time.Millisecond)
+		}
+	}
+	close(finish)
+	deadline = time.After(2 * time.Second)
+	for atomic.LoadInt32(&active) > 0 {
+		select {
+		case <-deadline:
+			t.Fatal("jobs did not drain")
+		default:
+			time.Sleep(time.Millisecond)
+		}
+	}
+	if got := atomic.LoadInt32(&maxActive); got != 1 {
+		t.Fatalf("expected max concurrency 1, got %d", got)
+	}
+}
+
+func TestDifferentServicesRunInParallel(t *testing.T) {
+	q := New()
+	started := make(chan struct{}, 2)
+	release := make(chan struct{})
+	for i := 0; i < 2; i++ {
+		q.Enqueue(uuid.New(), Job{Run: func(ctx context.Context) {
+			started <- struct{}{}
+			<-release
+		}})
+	}
+	for i := 0; i < 2; i++ {
+		select {
+		case <-started:
+		case <-time.After(2 * time.Second):
+			t.Fatal("parallel services did not both start")
+		}
+	}
+	close(release)
+}
+
+func TestCancelQueuedRemovesJob(t *testing.T) {
+	q := New()
+	svc := uuid.New()
+	release := make(chan struct{})
+	q.Enqueue(svc, Job{DeploymentID: uuid.New(), Run: func(ctx context.Context) { <-release }})
+
+	queuedID := uuid.New()
+	ran := make(chan struct{})
+	q.Enqueue(svc, Job{DeploymentID: queuedID, Run: func(ctx context.Context) { close(ran) }})
+
+	if got := q.Cancel(svc, queuedID); got != WasQueued {
+		t.Fatalf("expected WasQueued, got %v", got)
+	}
+	close(release)
+	select {
+	case <-ran:
+		t.Fatal("cancelled queued job still ran")
+	case <-time.After(200 * time.Millisecond):
+	}
+}
+
+func TestCancelActiveCancelsContext(t *testing.T) {
+	q := New()
+	svc := uuid.New()
+	id := uuid.New()
+	ctxDone := make(chan struct{})
+	q.Enqueue(svc, Job{DeploymentID: id, Run: func(ctx context.Context) {
+		<-ctx.Done()
+		close(ctxDone)
+	}})
+
+	deadline := time.After(2 * time.Second)
+	for {
+		q.mu.Lock()
+		_, ok := q.cancels[id]
+		q.mu.Unlock()
+		if ok {
+			break
+		}
+		select {
+		case <-deadline:
+			t.Fatal("job never registered")
+		default:
+			time.Sleep(time.Millisecond)
+		}
+	}
+	if got := q.Cancel(svc, id); got != WasActive {
+		t.Fatalf("expected WasActive, got %v", got)
+	}
+	select {
+	case <-ctxDone:
+	case <-time.After(2 * time.Second):
+		t.Fatal("active job context not cancelled")
+	}
+}
+
+func TestEnqueuePosition(t *testing.T) {
+	q := New()
+	svc := uuid.New()
+	release := make(chan struct{})
+	if pos := q.Enqueue(svc, Job{Run: func(ctx context.Context) { <-release }}); pos != 0 {
+		t.Fatalf("first job position = %d, want 0", pos)
+	}
+	if pos := q.Enqueue(svc, Job{Run: func(ctx context.Context) {}}); pos != 1 {
+		t.Fatalf("second job position = %d, want 1", pos)
+	}
+	if pos := q.Enqueue(svc, Job{Run: func(ctx context.Context) {}}); pos != 2 {
+		t.Fatalf("third job position = %d, want 2", pos)
+	}
+	close(release)
+}

app/backend/internal/docker/client.go

@@ -297,6 +297,7 @@ func (c *Client) BuildImage(ctx context.Context, buildContext io.Reader, options
 		BuildArgs:  options.BuildArgs,
 		Labels:     options.Labels,
 		Remove:     options.Remove,
+		NoCache:    options.NoCache,
 	})
 }
 

app/backend/internal/docker/types.go

@@ -44,6 +44,7 @@ type BuildOptions struct {
 	BuildArgs  map[string]*string
 	Labels     map[string]string
 	Remove     bool
+	NoCache    bool
 }
 
 // NetworkConfig represents the configuration for creating a network

app/backend/internal/mcpserver/server.go

@@ -64,15 +64,16 @@ func init() {
 
 	// Variables
 	register(spec{"containr_variables_list", "List service environment variables (secrets masked)", "GET", "/services/{id}/variables", []string{"id"}, nil, false})
-	register(spec{"containr_variables_set", "Replace all variables on a service. Args: id, variables = [{key, value, is_secret}]. Fetch the current list first and merge — the server replaces wholesale.", "PUT", "/services/{id}/variables", []string{"id"}, []string{"variables"}, false})
+	register(spec{"containr_variables_set", "Replace all variables on a service. Args: id, variables = [{key, value, is_secret}], redeploy (queue a redeploy to apply). Fetch the current list first and merge — the server replaces wholesale.", "PUT", "/services/{id}/variables", []string{"id"}, []string{"variables", "redeploy"}, false})
 
 	// Deployments
-	register(spec{"containr_deploy", "Trigger a deployment for a service. Args: id (service), commit_hash, branch.", "POST", "/services/{id}/deployments", []string{"id"}, []string{"commit_hash", "branch", "trigger"}, false})
+	register(spec{"containr_deploy", "Trigger a deployment for a service. Args: id (service), commit_hash, branch, no_cache.", "POST", "/services/{id}/deployments", []string{"id"}, []string{"commit_hash", "branch", "trigger", "no_cache"}, false})
 	register(spec{"containr_deployments_list", "List recent deployments across services", "GET", "/deployments", nil, nil, false})
 	register(spec{"containr_services_deployments", "List deployments for one service", "GET", "/services/{id}/deployments", []string{"id"}, nil, false})
 	register(spec{"containr_deployments_get", "Get a deployment", "GET", "/deployments/{id}", []string{"id"}, nil, false})
 	register(spec{"containr_deployments_logs", "Get persisted build/runtime logs for a deployment. Args: id, type (all|build|runtime).", "GET", "/deployments/{id}/logs", []string{"id"}, nil, false})
 	register(spec{"containr_deployments_rollback", "Roll back to a deployment. Requires confirm:true.", "POST", "/deployments/{id}/rollback", []string{"id"}, nil, true})
+	register(spec{"containr_deployments_cancel", "Cancel a queued or running deployment", "POST", "/deployments/{id}/cancel", []string{"id"}, nil, false})
 
 	// Databases
 	register(spec{"containr_databases_list", "List managed databases", "GET", "/databases", nil, nil, false})

app/backend/internal/types/build.go

@@ -28,6 +28,7 @@ type BuildRequest struct {
 	TriggeredBy  string `json:"triggered_by"`  // Who triggered the build
 	Branch       string `json:"branch"`        // Git branch
 	Commit       string `json:"commit"`        // Git commit SHA
+	NoCache      bool   `json:"no_cache"`      // Disable Docker layer cache
 }
 
 // BuildResponse represents the response from a build operation

app/frontend/src/features/service/pages/ServiceDetailPage.tsx

@@ -28,6 +28,7 @@ import {
   manualScaleService,
   execInService,
   rollbackDeployment,
+  cancelDeployment,
   updateServiceVariables,
   updateService,
   restartService,
@@ -66,6 +67,7 @@ import {
   Sparkles,
   RefreshCw,
   Box,
+  Ban,
   Play,
   ChevronDown,
   GitPullRequest,
@@ -104,7 +106,9 @@ function StatusBadge({ status }: { status: string }) {
     failed: { color: 'var(--error)', bg: 'var(--error-soft)', Icon: X, animate: false },
     building: { color: 'var(--warning)', bg: 'var(--warning-soft)', Icon: Loader2, animate: true },
     pending: { color: 'var(--warning)', bg: 'var(--warning-soft)', Icon: Loader2, animate: true },
+    queued: { color: 'var(--warning)', bg: 'var(--warning-soft)', Icon: Clock, animate: false },
     rolling_back: { color: 'var(--warning)', bg: 'var(--warning-soft)', Icon: RefreshCw, animate: true },
+    cancelled: { color: 'var(--text-tertiary)', bg: 'var(--surface-muted)', Icon: Ban, animate: false },
     stopped: { color: 'var(--text-tertiary)', bg: 'var(--surface-muted)', Icon: Box, animate: false },
   }[status] || { color: 'var(--text-tertiary)', bg: 'var(--surface-muted)', Icon: Box, animate: false };
 
@@ -374,6 +378,14 @@ export function ServiceDetailPage() {
     },
   });
 
+  const cancelDeploymentMutation = useMutation({
+    mutationFn: (deploymentId: string) => cancelDeployment(deploymentId),
+    onSuccess: () => {
+      queryClient.invalidateQueries({ queryKey: ['service-deployments', serviceId] });
+      queryClient.invalidateQueries({ queryKey: ['service', serviceId] });
+    },
+  });
+
   const rollbackMutation = useMutation({
     mutationFn: (deploymentId: string) => rollbackDeployment(deploymentId),
     onSuccess: () => {
@@ -514,7 +526,7 @@ export function ServiceDetailPage() {
     const deployments = isDemoMode ? [] : deploymentsQuery.data ?? [];
     const total = deployments.length;
     const active = deployments.filter((deployment) =>
-      ['pending', 'building', 'deploying', 'rolling_back'].includes(deployment.status),
+      ['queued', 'pending', 'building', 'deploying', 'rolling_back'].includes(deployment.status),
     ).length;
     const failed = deployments.filter((deployment) => deployment.status === 'failed').length;
 
@@ -2121,6 +2133,19 @@ export function ServiceDetailPage() {
                                 >
                                   Logs
                                 </button>
+                                {['queued', 'pending', 'building', 'deploying', 'rolling_back'].includes(deployment.status) && (
+                                  <button
+                                    onClick={() => {
+                                      if (window.confirm(`Cancel deployment ${deployment.id.slice(0, 8)}?`)) {
+                                        cancelDeploymentMutation.mutate(deployment.id);
+                                      }
+                                    }}
+                                    disabled={cancelDeploymentMutation.isPending}
+                                    className="px-3 py-1.5 rounded-[var(--radius-sm)] border border-[var(--border-subtle)] text-xs font-medium text-[var(--error)] hover:border-[var(--error)]/50 disabled:opacity-50 transition-colors"
+                                  >
+                                    {cancelDeploymentMutation.isPending && cancelDeploymentMutation.variables === deployment.id ? '...' : 'Cancel'}
+                                  </button>
+                                )}
                                 <button
                                   onClick={() => {
                                     if (window.confirm(`Rollback to deployment ${deployment.id.slice(0, 8)}?`)) {

app/frontend/src/generated/api-types.ts

@@ -1146,6 +1146,72 @@ export interface paths {
         patch?: never;
         trace?: never;
     };
+    "/deployments/{id}/cancel": {
+        parameters: {
+            query?: never;
+            header?: never;
+            path?: never;
+            cookie?: never;
+        };
+        get?: never;
+        put?: never;
+        /**
+         * Cancel a deployment
+         * @description Cancel a queued or in-flight deployment. Queued deployments are marked cancelled immediately; running ones get their build/reconcile aborted and settle as 'cancelled' shortly after.
+         */
+        post: {
+            parameters: {
+                query?: never;
+                header?: never;
+                path: {
+                    /** @description Deployment ID */
+                    id: string;
+                };
+                cookie?: never;
+            };
+            requestBody?: never;
+            responses: {
+                /** @description Queued deployment cancelled */
+                200: {
+                    headers: {
+                        [name: string]: unknown;
+                    };
+                    content: {
+                        "application/json": {
+                            /** @example cancelled */
+                            status?: string;
+                        };
+                    };
+                };
+                /** @description Running deployment is cancelling */
+                202: {
+                    headers: {
+                        [name: string]: unknown;
+                    };
+                    content: {
+                        "application/json": {
+                            /** @example cancelling */
+                            status?: string;
+                        };
+                    };
+                };
+                /** @description Deployment already finished */
+                409: {
+                    headers: {
+                        [name: string]: unknown;
+                    };
+                    content: {
+                        "application/json": components["schemas"]["Error"];
+                    };
+                };
+            };
+        };
+        delete?: never;
+        options?: never;
+        head?: never;
+        patch?: never;
+        trace?: never;
+    };
     "/deployments/{id}/rollback": {
         parameters: {
             query?: never;
@@ -7796,6 +7862,8 @@ export interface components {
             env_vars?: {
                 [key: string]: string;
             };
+            /** @description Build without Docker layer cache */
+            no_cache?: boolean;
         };
         RollbackDeploymentResponse: {
             deployment?: components["schemas"]["Deployment"];
@@ -8347,6 +8415,8 @@ export interface components {
         };
         UpdateVariablesRequest: {
             variables: components["schemas"]["VariableInput"][];
+            /** @description Queue a service redeploy after saving so running containers pick up the new variables */
+            redeploy?: boolean;
         };
         ConnectGitHubAppRequest: {
             /**

app/frontend/src/generated/docs-snapshot.ts

@@ -40,7 +40,7 @@ export const bundledDocs: BundledDoc[] = [
     "path": "guides/CLI.md",
     "section": "guides",
     "title": "Containr CLI",
-    "markdown": "# Containr CLI\n\nThe CLI controls every part of a Containr instance — the same API surface\nthe web UI uses. Build it with `scripts/build-cli.sh` or\n`go build ./cmd/cli` from `app/backend`.\n\n## Authentication\n\nCreate a personal access token in the web UI under\n**Settings → Personal Access Tokens** (or `containr tokens create`), then:\n\n```bash\ncontainr auth login cnp_...                      # verify + store\ncontainr auth login cnp_... --url https://containr.example.com\ncontainr auth login cnp_... --name production    # named profile\ncontainr auth status                             # show active profile + identity\ncontainr auth logout\n```\n\nResolution order for every command: `--token`/`--api-url` flags >\n`CONTAINR_TOKEN`/`CONTAINR_API_URL` env > active profile >\n`http://localhost:8080/api/v1`. The active profile is `--profile` >\n`CONTAINR_PROFILE` > `current_profile` in `~/.containr.yaml`.\n\n## Output and scripting\n\n- `--json` prints the raw API response — safe to pipe and parse.\n- `--yes` / `-y` skips confirmation on destructive commands (delete,\n  revoke, restore, failover). Non-interactive shells refuse destructive\n  commands without it.\n- Exit codes are stable: `0` ok, `1` error, `2` auth/permission,\n  `3` not found, `4` validation, `5` unreachable.\n\n## Commands\n\n```\nauth login|logout|status         token auth and profiles\ntokens list|create|revoke        personal access tokens\nprojects list|get|create|delete\nservices list|get|create|delete|start|stop|restart|redeploy\ndeploy <service-id>              trigger a deployment\ndeployments list|get|logs|rollback\nlogs <service-id> [--follow]     runtime logs (SSE follow)\nexec <service-id> -- <cmd>       one-off container command\nvariables list|set|unset         env vars (secrets masked)\ndatabases list|get|create|delete|action|backup|restore|download-backup\ncron list|get|create|delete|trigger|executions\ntemplates list|get|create|delete|deploy\nnodes list|get|delete|tokens     node agents + onboarding tokens (admin)\nscaling status|policies|services|scale\nha status|alerts|health|policies|enable|disable|failover\nsecurity scan|vulnerabilities|metrics\ngateway services|keys|analytics  (admin)\nadmin overview|users|settings|audit-logs\nnotifications list|read\nup                               deploy the current directory\n```\n\n## containr up\n\nRun inside a git checkout: uses `remote.origin.url` + the current branch\nas the build source, creates a service named after the directory when none\nexists, then triggers a deployment.\n\n```bash\ncontainr up --project <id>\ncontainr up --project <id> --image ghcr.io/example/app:latest   # image, no build\n```\n\nSet `project_id` on the active profile to omit `--project` (stored in\n`~/.containr.yaml` under `profiles.<name>.project_id`).\n\n## MCP server\n\n`containr-mcp` (`go build ./cmd/mcp`) exposes the same API as MCP tools\nover stdio — every command group above maps to a `containr_*` tool\n(~60 tools). Destructive tools require the caller to pass\n`confirm: true`; results are pretty JSON with `isError` on failures.\n\nClient configuration (Claude Code, Cursor, Devin, any stdio MCP host):\n\n```json\n{\n  \"mcpServers\": {\n    \"containr\": {\n      \"command\": \"containr-mcp\",\n      \"env\": {\n        \"CONTAINR_API_URL\": \"https://containr.example.com\",\n        \"CONTAINR_TOKEN\": \"cnp_...\"\n      }\n    }\n  }\n}\n```\n\nScope the token deliberately: `read` is enough for inspection agents,\n`write` for deploy automation, `admin` only for platform operations.\n"
+    "markdown": "# Containr CLI\n\nThe CLI controls every part of a Containr instance — the same API surface\nthe web UI uses. Build it with `scripts/build-cli.sh` or\n`go build ./cmd/cli` from `app/backend`.\n\n## Authentication\n\nCreate a personal access token in the web UI under\n**Settings → Personal Access Tokens** (or `containr tokens create`), then:\n\n```bash\ncontainr auth login cnp_...                      # verify + store\ncontainr auth login cnp_... --url https://containr.example.com\ncontainr auth login cnp_... --name production    # named profile\ncontainr auth status                             # show active profile + identity\ncontainr auth logout\n```\n\nResolution order for every command: `--token`/`--api-url` flags >\n`CONTAINR_TOKEN`/`CONTAINR_API_URL` env > active profile >\n`http://localhost:8080/api/v1`. The active profile is `--profile` >\n`CONTAINR_PROFILE` > `current_profile` in `~/.containr.yaml`.\n\n## Output and scripting\n\n- `--json` prints the raw API response — safe to pipe and parse.\n- `--yes` / `-y` skips confirmation on destructive commands (delete,\n  revoke, restore, failover). Non-interactive shells refuse destructive\n  commands without it.\n- Exit codes are stable: `0` ok, `1` error, `2` auth/permission,\n  `3` not found, `4` validation, `5` unreachable.\n\n## Commands\n\n```\nauth login|logout|status         token auth and profiles\ntokens list|create|revoke        personal access tokens\nprojects list|get|create|delete\nservices list|get|create|delete|start|stop|restart|redeploy\ndeploy <service-id>              trigger a deployment (--no-cache, --commit-hash, --branch)\ndeployments list|get|logs|rollback|cancel\nlogs <service-id> [--follow]     runtime logs (SSE follow)\nexec <service-id> -- <cmd>       one-off container command\nvariables list|set|unset         env vars (secrets masked; --redeploy applies them)\ndatabases list|get|create|delete|action|backup|restore|download-backup\ncron list|get|create|delete|trigger|executions\ntemplates list|get|create|delete|deploy\nnodes list|get|delete|tokens     node agents + onboarding tokens (admin)\nscaling status|policies|services|scale\nha status|alerts|health|policies|enable|disable|failover\nsecurity scan|vulnerabilities|metrics\ngateway services|keys|analytics  (admin)\nadmin overview|users|settings|audit-logs\nnotifications list|read\nup                               deploy the current directory\n```\n\n## containr up\n\nRun inside a git checkout: uses `remote.origin.url` + the current branch\nas the build source, creates a service named after the directory when none\nexists, then triggers a deployment.\n\n```bash\ncontainr up --project <id>\ncontainr up --project <id> --image ghcr.io/example/app:latest   # image, no build\n```\n\nSet `project_id` on the active profile to omit `--project` (stored in\n`~/.containr.yaml` under `profiles.<name>.project_id`).\n\n## MCP server\n\n`containr-mcp` (`go build ./cmd/mcp`) exposes the same API as MCP tools\nover stdio — every command group above maps to a `containr_*` tool\n(~60 tools). Destructive tools require the caller to pass\n`confirm: true`; results are pretty JSON with `isError` on failures.\n\nClient configuration (Claude Code, Cursor, Devin, any stdio MCP host):\n\n```json\n{\n  \"mcpServers\": {\n    \"containr\": {\n      \"command\": \"containr-mcp\",\n      \"env\": {\n        \"CONTAINR_API_URL\": \"https://containr.example.com\",\n        \"CONTAINR_TOKEN\": \"cnp_...\"\n      }\n    }\n  }\n}\n```\n\nScope the token deliberately: `read` is enough for inspection agents,\n`write` for deploy automation, `admin` only for platform operations.\n\nDeployments are serialized per service: a second `deploy` on a busy\nservice lands in `queued` state and runs when the active one finishes.\n`deployments cancel <id>` drops a queued deployment or aborts a running\none (builds/reconciles stop via context cancellation).\n"
   },
   {
     "path": "guides/CLOUDFLARE_SETUP.md",

app/frontend/src/lib/api-client.ts

@@ -1320,6 +1320,12 @@ export async function rollbackDeployment(deploymentId: string): Promise<Rollback
   };
 }
 
+export async function cancelDeployment(deploymentId: string): Promise<{ status: string }> {
+  return requestJson<{ status: string }>(`/deployments/${deploymentId}/cancel`, {
+    method: 'POST',
+  });
+}
+
 export async function listBuilds(input: ListBuildsInput = {}): Promise<ListBuildsResult> {
   const searchParams = new URLSearchParams();
   if (input.projectId) {

docs/api/openapi.yaml

@@ -811,6 +811,45 @@ paths:
                   deployment:
                     $ref: '#/components/schemas/Deployment'
 
+  /deployments/{id}/cancel:
+    post:
+      tags:
+        - Deployments
+      summary: Cancel a deployment
+      description: Cancel a queued or in-flight deployment. Queued deployments are marked cancelled immediately; running ones get their build/reconcile aborted and settle as 'cancelled' shortly after.
+      security:
+        - bearerAuth: []
+      parameters:
+        - name: id
+          in: path
+          required: true
+          description: Deployment ID
+          schema:
+            type: string
+      responses:
+        '200':
+          description: Queued deployment cancelled
+          content:
+            application/json:
+              schema:
+                type: object
+                properties:
+                  status: { type: string, example: cancelled }
+        '202':
+          description: Running deployment is cancelling
+          content:
+            application/json:
+              schema:
+                type: object
+                properties:
+                  status: { type: string, example: cancelling }
+        '409':
+          description: Deployment already finished
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/Error'
+
   /deployments/{id}/rollback:
     post:
       tags:
@@ -5108,6 +5147,9 @@ components:
           additionalProperties:
             type: string
           description: Runtime environment variable overrides
+        no_cache:
+          type: boolean
+          description: Build without Docker layer cache
 
     RollbackDeploymentResponse:
       type: object
@@ -5989,6 +6031,9 @@ components:
           type: array
           items:
             $ref: '#/components/schemas/VariableInput'
+        redeploy:
+          type: boolean
+          description: Queue a service redeploy after saving so running containers pick up the new variables
 
     ConnectGitHubAppRequest:
       type: object

docs/guides/CLI.md

@@ -38,11 +38,11 @@ auth login|logout|status         token auth and profiles
 tokens list|create|revoke        personal access tokens
 projects list|get|create|delete
 services list|get|create|delete|start|stop|restart|redeploy
-deploy <service-id>              trigger a deployment
-deployments list|get|logs|rollback
+deploy <service-id>              trigger a deployment (--no-cache, --commit-hash, --branch)
+deployments list|get|logs|rollback|cancel
 logs <service-id> [--follow]     runtime logs (SSE follow)
 exec <service-id> -- <cmd>       one-off container command
-variables list|set|unset         env vars (secrets masked)
+variables list|set|unset         env vars (secrets masked; --redeploy applies them)
 databases list|get|create|delete|action|backup|restore|download-backup
 cron list|get|create|delete|trigger|executions
 templates list|get|create|delete|deploy
@@ -95,3 +95,8 @@ Client configuration (Claude Code, Cursor, Devin, any stdio MCP host):
 
 Scope the token deliberately: `read` is enough for inspection agents,
 `write` for deploy automation, `admin` only for platform operations.
+
+Deployments are serialized per service: a second `deploy` on a busy
+service lands in `queued` state and runs when the active one finishes.
+`deployments cancel <id>` drops a queued deployment or aborts a running
+one (builds/reconciles stop via context cancellation).

llms.txt

@@ -12,9 +12,10 @@ Base URL: `https://<host>/api/v1`. OpenAPI: `docs/api/openapi.yaml`.
 
 ## Conventions
 
-- **Errors**: `{"error": "msg", "code": "MACHINE_CODE"}`. Codes: `UNAUTHENTICATED`, `INVALID_TOKEN`, `FORBIDDEN`, `ADMIN_REQUIRED`, `SCOPE_INSUFFICIENT`, `NOT_FOUND`, `VALIDATION`, `CONFLICT`, `INTERNAL`, `DEPENDENCY_UNAVAILABLE`, `IDEMPOTENCY_IN_PROGRESS`.
+- **Errors**: `{"error": "msg", "code": "MACHINE_CODE"}`. Codes: `UNAUTHENTICATED`, `INVALID_TOKEN`, `FORBIDDEN`, `ADMIN_REQUIRED`, `SCOPE_INSUFFICIENT`, `NOT_FOUND`, `VALIDATION`, `CONFLICT`, `INTERNAL`, `DEPENDENCY_UNAVAILABLE`, `IDEMPOTENCY_IN_PROGRESS`, `NOT_CANCELLABLE`.
 - **Idempotency**: send `Idempotency-Key: <uuid>` on POSTs — the first response replays for 24h; concurrent duplicates get 409. Always safe to send.
 - **Pagination**: `?limit=` + `?cursor=` (pass `pagination.next_cursor` verbatim; legacy `?page=` also works). Invalid cursors fall back to page 1.
+- **Deployments**: serialized per service — a deploy on a busy service returns `status: queued` and runs when the active one finishes. `POST /deployments/{id}/cancel` drops queued (`200 cancelled`) or aborts running (`202 cancelling`) work. `no_cache: true` on create disables the Docker layer cache. `PUT /services/{id}/variables` accepts `redeploy: true` to apply env changes to running containers.
 - **Identifiers**: all ids are UUIDs. Tokens: `cnp_` user PATs, `cagt_` agent onboarding tokens.
 
 ## Core objects