DAG based swarm jobs for https://www.reddit.com/r/docker/comments/9511ks/how_should_i_setup_periodic_tasks_in_docker/

docker-compose.yml

---
## Execute: docker compose up -D ./docker-compose.yml
version: '3'
services:

  autoscale:
    image: gianarb/orbiter:latest
    healthcheck:
      test: ["CMD", "exit 0"]
      interval: 25s
      timeout: 5s
      retries: 3
    networks:
      default:
        aliases:
          - autoscale.local
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock:rw
    deploy:
      placement:
        constraints: [node.role == manager]
      mode: global
      restart_policy:
        condition: on-failure
        max_attempts: 3
        delay: 5s
      rollback_config:
        parallelism: 2
        delay: 10s
        monitor: 10s
        order: start-first
        failure_action: continue
      update_config:
        parallelism: 2
        delay: 10s
        monitor: 10s
        order: start-first
        failure_action: rollback
      labels:
        orbiter: true
    ports:
      - '8000/tcp'

  autoupgrade:
    image: mazzolino/shepherd
    environment:
      TZ: 'UTC'
      FILTER_SERVICES: ''
      IGNORELIST_SERVICES: ""
      RUN_ONCE_AND_EXIT: "true"
    healthcheck:
      test: ["CMD", "exit 0"]
      interval: 25s
      timeout: 5s
      retries: 3
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock:ro
    deploy:
      replicas: 0
      restart_policy:
        condition: none
      labels:
        - swarm.cronjob.enable=true
        # Start service every ten minutes to look for new images
        - swarm.cronjob.schedule=*/10 * * * * *
        - swarm.cronjob.skip-running=true
      placement:
        constraints:
        - node.role == manager

  swarm-cronjob:
    image: crazymax/swarm-cronjob
    healthcheck:
      test: ["CMD", "exit 0"]
      interval: 25s
      timeout: 5s
      retries: 3
    volumes:
      - "/var/run/docker.sock:/var/run/docker.sock:ro"
    environment:
      - "TZ=UTC"
      - "LOG_LEVEL=info"
      - "LOG_JSON=false"
    deploy:
      mode: global
      placement:
        constraints:
          - node.role == manager

  check_for_json:
    healthcheck:
      test: ["CMD", "exit 0"]
      interval: 25s
      timeout: 5s
      retries: 3
    deploy:
      mode: replicated
      replicas: 1
      labels:
        orbiter: false
        swarm.cronjob.enable: true
        swarm.cronjob.schedule: "0 */1 * * * *" # on every hour *:00:00
        swarm.cronjob.skip-running: false
      restart_policy:
        condition: any
        delay: 1h       ### << every hour after deploy
      volumes:
        - /srv/json:/data:ro
      image: alpine:latest
      command: >
        test -e /data/MySpecialFile.json && curl -vX POST autoscale.local:8000/v1/orbiter/handle/autoswarm/my_example_php_script/up
       depends_on:
         autoscaler: service_healthy

  my_example_php_script:
    healthcheck:
      test: ["CMD", "exit 0"]
      interval: 25s
      timeout: 5s
      retries: 3
    image: my_php_script:latest
    build:
      context: .
      args:
        SCRIPT_NAME=myPhpScript.php
      dockerfile_inline: |
        FROM 8.1-cli-alpine3.18
        ARG SCRIPT_NAME=index.php
        WORKDIR /src
        COPY ./$SCRIPT_NAME /src/
        WORKDIR /data
        RUN apk add mysql-client
        ENTRYPOINT /usr/bin/php
        CMD /src/$SCRIPT_NAME
    depends_on:
      autoscale: service_health
    environment:
      - MYSQL_HOST="example.com"
    volumes:
      - /srv/json:/data:ro
    deploy:
      restart_policy: none
      replicas: 0
      mode: replicated
      labels:
        orbiter: true
        orbiter.up: "1"
        orbiter.down: "0"
...