feat(scheduling): multi-node service placement + remote lifecycle (#54)

* feat(scheduling): multi-node service placement + remote container lifecycle

services.node_id pins a service to a node agent: explicit id, "auto"
(least-loaded online node), or "local"/empty for the local Docker host.

- DeploymentEngine gains a NodeRunner seam; ReconcileService routes
  pinned services to the agent's command queue (create_container with
  image/env/command/ports/volumes/labels/resource limits), tracked in
  container_instances with deterministic per-replica ids.
- Stop/start/restart/delete fan out to the pinned node — no more leaked
  remote containers on service delete.
- Agent create_container is idempotent (rm -f first) and honors restart
  policy, labels, memory/cpus, published+ephemeral ports, volumes, env.
- GET /nodes/options exposes id/name/status to authed users; full agent
  management stays admin-only.
- CLI services create/update --node, MCP node_id docs, OpenAPI
  node_id/node_name + /nodes/options, service-detail Placement section.

Ceilings: remote nodes run registry-pulled images only — git builds,
Traefik domains, and remote sleep detection stay local until
build-on-node and per-node ingress land.

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

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

* feat(scheduling): node cordon/drain + schedulable gate

- node_agents.schedulable column; "auto" placement and explicit pins skip
  cordoned nodes; already-pinned services keep running (reconcile is
  unaffected — cordon is not eviction).
- POST /agents/:id/cordon|uncordon|drain (admin). Drain cordons, removes
  every service container via the agent queue, and clears the services'
  pins so they redeploy on the local host. Requires the node online —
  an offline node can't acknowledge removals.
- CLI nodes cordon|uncordon|drain (drain confirm-gated), MCP
  containr_nodes_cordon|uncordon|drain, OpenAPI paths + schedulable
  field, node-detail page cordon badge + drain/cordon controls.
- Fix: re-registering an agent no longer resets auto_prune.

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

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

* feat(scheduling): remote registry pulls + post-run health check

- Dispatch payload carries the project owner's registry credentials for
  the image host; the agent docker-logins for the pull and logs out
  after, so private images deploy to remote nodes.
- Completed command payloads get their registry password scrubbed —
  secrets don't linger in agent_commands rows.
- Agent verifies the container is still running ~1.5s after docker run;
  crash-loops now fail the deployment with the container's last log
  lines instead of reporting "running" optimistically.

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

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

---------

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

app/backend/cmd/agent/main.go

@@ -283,7 +283,38 @@ func createContainer(payload map[string]interface{}) (string, error) {
 		return "", errors.New("container name and image are required")
 	}
 
-	args := []string{"run", "-d", "--name", name, "--restart", "unless-stopped"}
+	// Private registry: log in for the pull, then log back out so the
+	// credential doesn't linger in the node's docker config.
+	if reg, ok := container["registry"].(map[string]interface{}); ok {
+		if user := stringValue(reg["username"]); user != "" {
+			server := firstString(stringValue(reg["server"]), registryHostOf(image))
+			login := exec.Command("docker", "login", server, "-u", user, "--password-stdin")
+			login.Stdin = strings.NewReader(stringValue(reg["password"]))
+			if out, err := login.CombinedOutput(); err != nil {
+				return "", fmt.Errorf("registry login %s: %s", server, strings.TrimSpace(string(out)))
+			}
+			defer func() { _, _ = runDocker("logout", server) }()
+		}
+	}
+
+	// Redeploys reuse the deterministic name — clear the stale container
+	// first so `run` never fails on a name conflict.
+	_, _ = runDocker("rm", "-f", name)
+
+	args := []string{"run", "-d", "--name", name}
+	restart := firstString(stringValue(container["restart_policy"]), "unless-stopped")
+	args = append(args, "--restart", restart)
+	if labels, ok := container["labels"].(map[string]interface{}); ok {
+		for key, value := range labels {
+			args = append(args, "--label", fmt.Sprintf("%s=%v", key, value))
+		}
+	}
+	if mem := int64Value(container["memory"]); mem > 0 {
+		args = append(args, "--memory", fmt.Sprintf("%d", mem))
+	}
+	if cpus := floatValue(container["cpus"]); cpus > 0 {
+		args = append(args, "--cpus", fmt.Sprintf("%g", cpus))
+	}
 	if env, ok := container["environment"].(map[string]interface{}); ok {
 		for key, value := range env {
 			args = append(args, "-e", fmt.Sprintf("%s=%v", key, value))
@@ -298,8 +329,14 @@ func createContainer(payload map[string]interface{}) (string, error) {
 			hostPort := intValue(port["host_port"])
 			containerPort := intValue(port["container_port"])
 			protocol := firstString(stringValue(port["protocol"]), "tcp")
-			if hostPort > 0 && containerPort > 0 {
+			if containerPort <= 0 {
+				continue
+			}
+			if hostPort > 0 {
 				args = append(args, "-p", fmt.Sprintf("%d:%d/%s", hostPort, containerPort, protocol))
+			} else {
+				// Ephemeral host port — docker picks the binding.
+				args = append(args, "-p", fmt.Sprintf("%d/%s", containerPort, protocol))
 			}
 		}
 	}
@@ -321,7 +358,33 @@ func createContainer(payload map[string]interface{}) (string, error) {
 		}
 	}
 	args = append(args, image)
-	return runDocker(args...)
+	if cmd, ok := container["command"].([]interface{}); ok {
+		for _, arg := range cmd {
+			args = append(args, stringValue(arg))
+		}
+	}
+	if _, err := runDocker(args...); err != nil {
+		return "", err
+	}
+	// docker run -d returns before the entrypoint fails — a crash-looping
+	// service would be reported healthy. Give it a moment, then verify.
+	time.Sleep(1500 * time.Millisecond)
+	state, _ := runDocker("inspect", "-f", "{{.State.Status}}", name)
+	if state != "running" {
+		logs, _ := runDocker("logs", "--tail", "5", name)
+		return "", fmt.Errorf("container %s is %q after start: %s", name, state, logs)
+	}
+	return name, nil
+}
+
+// registryHostOf mirrors the server's registryHost: first path segment with
+// a "." or ":" is a registry host; bare names mean Docker Hub.
+func registryHostOf(imageRef string) string {
+	parts := strings.Split(imageRef, "/")
+	if len(parts) > 1 && (strings.Contains(parts[0], ".") || strings.Contains(parts[0], ":")) {
+		return parts[0]
+	}
+	return "docker.io"
 }
 
 func commandTarget(payload map[string]interface{}) string {
@@ -620,3 +683,35 @@ func boolValue(value interface{}) bool {
 	typed, _ := value.(bool)
 	return typed
 }
+
+func int64Value(value interface{}) int64 {
+	switch typed := value.(type) {
+	case int64:
+		return typed
+	case int:
+		return int64(typed)
+	case float64:
+		return int64(typed)
+	case json.Number:
+		out, _ := typed.Int64()
+		return out
+	default:
+		return 0
+	}
+}
+
+func floatValue(value interface{}) float64 {
+	switch typed := value.(type) {
+	case float64:
+		return typed
+	case int:
+		return float64(typed)
+	case int64:
+		return float64(typed)
+	case json.Number:
+		out, _ := typed.Float64()
+		return out
+	default:
+		return 0
+	}
+}