Supervise the engine's host: deep health, loop watchdog, one worker

The API image ran four uvicorn workers, and each one built a full flow
controller — four sets of MQTT subscriptions, cron ticks and webhooks.
Runs one worker now; scaling out is the worker split, not more processes.

Adds a loop-lag watchdog and a deep /utils/health/ that fails when the
event loop is wedged or Redis is unreachable, the two failure modes a
process-alive check never sees. Autoheal restarts on that signal, behind
a compose profile because it mounts the Docker socket.

The private user-seeding routes now need an explicit opt-in rather than
just ENVIRONMENT=local, so a deployment that kept the default never
exposes them.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011LF61rxW1FG5YCD2J9YqjY

docker/compose.dev.yml

@@ -77,6 +77,8 @@ services:
       SMTP_PORT: "1025"
       SMTP_TLS: "false"
       EMAILS_FROM_EMAIL: "noreply@fluksio.com"
+      # Test-only user seeding, needed by the Playwright suite. Dev stack only.
+      PRIVATE_API_ENABLED: "true"
 
   mailcatcher:
     image: schickling/mailcatcher

docker/compose.yml

@@ -154,16 +154,20 @@ services:
     volumes:
       - app-flow-data:/data
 
+    # Deep health: fails when the event loop is wedged or Redis is gone, not
+    # just when the process is dead. Autoheal restarts on unhealthy.
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:8000/api/v1/utils/health-check/"]
+      test: ["CMD", "curl", "-f", "http://localhost:8000/api/v1/utils/health/"]
       interval: 10s
       timeout: 5s
       retries: 5
+      start_period: 30s
 
     build:
       context: ..
       dockerfile: backend/Dockerfile
     labels:
+      - autoheal=true
       - traefik.enable=true
       - traefik.docker.network=proxy
       - traefik.constraint-label=proxy
@@ -212,6 +216,28 @@ services:
 
       - traefik.http.routers.${STACK_NAME?Variable not set}-frontend-http.middlewares=https-redirect
 
+  # Docker never restarts a merely *unhealthy* container on its own; autoheal
+  # closes that gap for the services labeled autoheal=true.
+  #
+  # Behind a profile because it needs the Docker socket, which is host-wide
+  # authority: on a machine that runs anything besides this stack, that is a
+  # deliberate operator decision. `make up` opts in; the dev stacks do not.
+  autoheal:
+    image: willfarrell/autoheal:latest
+    container_name: fluksio-autoheal
+    profiles: ["autoheal"]
+    restart: always
+    security_opt:
+      - no-new-privileges:true
+    networks:
+      - default
+    environment:
+      # Scoped by label, so it only ever restarts this stack's backend.
+      - AUTOHEAL_CONTAINER_LABEL=autoheal
+      - AUTOHEAL_INTERVAL=15
+    volumes:
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
 volumes:
   app-db-data:
   app-redis-data: