docker-compose file improvement

[sandstromviktor]

Hello Team Whale.
Hope everyone is doing good :) I was doing some docker compose stuff and realized that there some cool yaml things that can simplify your docker-compose.yaml file (and of course other yaml files, like the charts etc).

When you define the studio service here

  studio:
    user: "${UID}:${GID}"
    container_name: studio
    ... etc etc...

you can define a yaml anchor studio: &studio
and then merge the config on other services using the yaml alias <<: *studio

Example

  studio: &studio # Define the anchor here
    user: "${UID}:${GID}"
    container_name: studio
    build:
      context: .
      target: runtime
    image: stackn:develop
    env_file:
      - .env
    command: ["scripts/wait-for-it.sh", "db:5432", "--", sh, scripts/run_web.sh]
    ports:
      - "8080:8080"
    volumes:
      - .:/app:cached
     # - ${PWD}/cluster.conf:/app/cluster.conf

# Now in the celery worker

  celery-worker:
    <<: *studio # Merge configs from studio using the alias
    # Note that i have removed a bunch of duplicated fields here
    container_name: celery-worker
    command: sh ./scripts/run_worker.sh
    volumes:
      - ${PWD}/cluster.conf:/app/cluster.conf
    links:
      - db
      - studio
      - rabbit
      - redis
    depends_on:
      - db
      - studio
      - rabbit
      - redis
    networks:
      internal_network:
        aliases:
          - celery.127.0.0.1.nip.io

What happens?

The alias merges all the config from studio and then you override specific fields afterwards. This makes it easier to define the user, image, env_file etc.. once in the file, instead of doing it multiple times for the different services.

All the best
/Viktor

PS. Serve looks awesome now:)

[alfredeen]

Hi Viktor, It is super-nice to hear from you and thank you so much for this great tip! It looks very useful and something I think we should make use of.