Setting Up Traefik Proxy with PostgreSQL and pgAdmin in Docker Compose

• How much do you know "Traefik" proxy?
• Traefik + Docker Compose

I'll share the Docker Compose file, followed by a full analysis of what every line in our template represents.

• Replace localhost with your own domain or sub-domain in traefik, postgres & pgadmin services
• Change config vars against these environment vars in postgres service — POSTGRES_USER, POSTGRES_PASSWORD & POSTGRES_DB
• Change config vars against these environment vars in pgadmin service — PGADMIN_DEFAULT_EMAIL, PGADMIN_DEFAULT_PASSWORD & SCRIPT_NAME
• Run docker-compose.yml file by using the command docker-compose up -d
• Your pgAdmin server should be up and running. Visiting http://localhost:1337/_admin/pgadmin, will show you your pgAdmin portal
• You'll need to register a server in pgAdmin after you successfully login into it with the configs values you used in your docker-compose.yml file - just remember to use pg_container as the hostname

• I'm using official docker images for PostgreSQL & pgAdmin

• The container name is your database hostname, from our template the database hostname is pg_container
• depends_on: [traefik] means we want to wait till traefik service is up
• ports exposes our postgres to accept incoming requests with the postgres port running inside your docker container
• volumes mapping helps us to keep our data within the database to persist in our local machine. So, next time when we run our docker-compose from the same location - persisted data from the directory ./postgres-data will be utilised by our postgres service container and it won't redo all the installation steps
• In volumes, you see I have commented out this line — # - ./init.sql:/docker-entrypoint-initdb.d/init.sql
  • If you would like to do additional initialization in an image derived from this one, add one or more *.sql, *.sql.gz, or *.sh scripts under /docker-entrypoint-initdb.d (creating the directory if necessary)
  • After the entrypoint calls initdb to create the default postgres user and database, it will run any *.sql files, run any executable *.sh scripts, and source any non-executable *.sh scripts found in that directory to do further initialization before starting the service
  • Warning: scripts in /docker-entrypoint-initdb.d are only run if you start the container with a data directory that is empty; any pre-existing database will be left untouched on container startup
  • One common problem is that if one of your /docker-entrypoint-initdb.d scripts fails (which will cause the entrypoint script to exit) and your orchestrator restarts the container with the already initialized data directory, it will not continue on with your scripts
• There are lots of other environment vars available in postgres docker image, you can find their references here - postgres
• I do healthcheck with postgres container because it takes almost a minute to initialise for the first time and starts accepting connection requests once postgres container is up & this might break my backend apps which would need database connection in their start scripts

• You can find healthcheck references from the docker-compose's official documentation

• depends_on: postgres This statement means that pgAdmin service depends on postgres and waits until the healthy service status is received from the same container and this is how one should use the postgres depends in their backend app services as well.
• If not done as above, your backend will definitely fail most of the times which is not good for your final product
• Almost all of the configs speaks for itself except env var SCRIPT_NAME
• Since I have traefik - I wanted to host my pgAdmin app on some route & here's how it's done
  • Firstly, to do this I had to move pgadmin on some route & you can do it by setting the SCRIPT_NAME variable

  • Once that's done, now rewriting routes uri using traefik labels to /_admin/pgadmin will do the job

And I strongly advise everyone to have the "health check" docker-compose's index in your postgres service container to avoid unnecessary initialisation failures.

Source: This blog is authored by Faiz Ahmed, Principal Technical Consultant at GeekyAnts. Originally published on Hashnode: Read here.