stack.yml

services:
  controller:
    image: ${CONTROLLER_IMAGE:?CONTROLLER_IMAGE must be set — run via deploy.sh or export it first}
    ports:
      - "${UI_PORT:-8080}:8080"
    volumes:
      - "${CONTROLLER_ROOT:-/opt/jenkins_home}:/var/jenkins_home"
    healthcheck:
      test: ["CMD-SHELL", "curl -fsS http://localhost:8080/jenkins/login >/dev/null || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 5
      start_period: 60s
    secrets:
      - jenkins-user
      - jenkins-pass
    deploy:
      replicas: 1
      restart_policy:
        condition: on-failure

  worker:
    user: root
    image: ${WORKER_IMAGE:?WORKER_IMAGE must be set — run via deploy.sh or export it first}
    environment:
      - JENKINS_CONTROLLER_URL=${JENKINS_CONTROLLER_URL:?JENKINS_CONTROLLER_URL must be set — run via deploy.sh or export it first}
      - WORKER_ROOT=${WORKER_ROOT:-/opt/worker_home}
      - SWARM_EXECUTORS=${SWARM_EXECUTORS:-5}
      - SWARM_LABELS=${SWARM_LABELS:-swarm docker}
      - SWARM_WEBSOCKET=${SWARM_WEBSOCKET:-true}
    volumes:
      - "${WORKER_ROOT:-/opt/worker_home}:${WORKER_ROOT:-/opt/worker_home}"
      - "/var/run/docker.sock:/var/run/docker.sock"
    secrets:
      - jenkins-user
      - jenkins-pass
    deploy:
      replicas: ${WORKER_REPLICAS:-1}
      # Retry window (~100s) must exceed controller start_period (60s) + one healthcheck interval (30s).
      # Increase max_attempts if the controller is slow to pass its healthcheck.
      restart_policy:
        condition: on-failure
        delay: 10s
        max_attempts: 10

secrets:
  jenkins-user:
    external: true
  jenkins-pass:
    external: true