Docker Compose

Almost nothing real is just one container. A typical app is a web server, an application process, and a database, all talking to each other. Docker Compose lets you describe all of them, and how they connect, in a single YAML file.

Anatomy of a compose file

services:
  web:
    image: nginx
    ports:
      - "80:80"
  app:
    build: .
    depends_on:
      - db
  db:
    image: postgres
    volumes:
      - db-data:/var/lib/postgresql/data

volumes:
  db-data:

• services - one entry per container: which image to run (or build), ports to expose, what it depends on.
• depends_on - start order; app won't start until db has.
• volumes - named, persistent storage that survives a container being recreated - essential for a database, which should never lose its data just because its container got rebuilt.

One command instead of many

Without Compose, standing up that same three-container app means three separate docker run commands, each with the right flags, in the right order, every single time. With a compose file written once:

docker compose up -d     # start everything
docker compose down      # stop and remove everything
docker compose logs -f   # follow logs from every service at once

Why this matters for local development

Compose is most commonly reached for to spin up an entire application stack - with its database, cache, and message queue - on a single laptop, in one command, identically for every developer on the team. No more "works on my machine because I have Postgres 14 and you have 15."

Try it

touch docker-compose.yml

When you're ready, hit Mark complete and move to the next lesson.