Set up separated development and production environment with docker-compose

In this post, I will walk through a quick example to summarize how to separate the development and production environment configurations in docker-compose.

Principle: How to separate development and production docker-compose.yml?

What we can do is to separate the different configurations into 3 files.

• docker-compose.yml
  This captures the common configuration needed in both environments
• docker-compose.override.yml
  This captures the configuration only needed in development environment. When executing docker-compose build or docker-compose up, docker-compose will automatically take the configurations from docker-compose.override.yml. We can continue to use the basic commands we use in development.
• docker-compose.prod.yml
  This captures the configuration only needed in production environment. We use -f to tell docker-compose how to link the configuration in production environment, like this:

docker-compose -f docker-compose.yml -f docker-compose.prod.yml up

Quick example

Let’s assume we originally have a docker-compose.yml which we use to set up things for development environment.

version: "3"
 
services:
    webapp:
        build: ./
        container_name: webapp
        restart: unless-stopped
        networks:
            - frontend

    nginx:
        build: ./nginx/
        container_name: webserver
        restart: unless-stopped
        volumes:
            - ./DockerLog/nginx:/var/log/nginx
        depends_on:
            - webapp
        ports:
            - "443:443"
            - "80:80"
        networks:
            - frontend

networks:
    frontend:
        driver: bridge

Now at some point, we want to deploy on a production server. The production server is using reverse proxy on an external network. We no longer need the frontend network and also don’t need the nginx there. To summarize the difference in our targeted development and production environments:

• In the development environment, we want to use frontend network which is a bridge. And in production environment, we want to use an external network.
• In the development environment, we want to use nginx. In production environment, we don’t need nginx. However, in the production environment, we need to add the URL and LETSENCRYPT configurations for the external nginx-proxy.

The common part of the configuration goes to docker-compose.yml:

version: "3"
 
services:
    webapp:
        build: ./
        container_name: webapp
        restart: unless-stopped

The rest of development configurations goes to docker-compose.override.yml:

version: "3"
 
services:
    webapp:
        networks:
            - frontend

    nginx:
        build: ./nginx/
        container_name: webserver
        restart: unless-stopped
        volumes:
            - ./DockerLog/nginx:/var/log/nginx
        depends_on:
            - webapp
        ports:
            - "443:443"
            - "80:80"
        networks:
            - frontend

networks:
    frontend:
        driver: bridge

The new configurations necessary for production environments goes to docker-compose.prod.yml

version: "3"
 
services:
    webapp:
        networks:
            - default
        environment:
            - VIRTUAL_HOST=<your url, e.g. www.company.com>
            - LETSENCRYPT_HOST=<your url, e.g. www.company.com>
            - LETSENCRYPT_EMAIL=<your email>

networks:
    default:
        external:
            name: nginx-proxy

We are assuming the external nginx-proxy network has been created here. If you are interested in learning more about nginx-proxy and how to set up docker networks, check this post for a walkthrough about the concepts.

Now, in development environment we bring up the web app by:

docker-compose up --build

In production environment, we bring up the web app by:

docker-compose -f docker-compose.yml -f docker-compose.prod.yml up

Reference

https://docs.docker.com/compose/extends/#different-environments