Skip to main content

Debugging in production

There are several failure points that we can look into to try and find the cause of a bug.

Deployment logs​

If a problem happens during deployment, you can check the status of the GitHub Action responsible for deploying the latest pull request to production here. This should provide enough informations for deployment related issues.

The deployment pipeline is split into two parts:

  1. Building the containers swarm.
  2. Running the containers.

We will see how to debug both processes.

info

In the following section, all docker-compose commands in production are replaced with the alias dcf which stands for docker-compose -f docker-compose.yml -f docker-compose.prod.yml

Errors while building​

The build logs should be available in the GitHub Action logs like described above. If you need more informations, re-run the build directly on the server:

  1. Connect to the server using SSH.
  2. Go to nantralPlatform/deployment by running cd nantralPlatform/deployment.
  3. Run sudo dcf build to re-build the containers.

Errors while running the containers​

Even if the docker-compose build ran successfuly, containers can still be prone to errors. If an error happens, you can debug the swarm by following the steps below:

  1. Connect to the server using SSH.
  2. Go to nantralPlatform/deployment by running cd nantralPlatform/deployment.
  3. Run dcf logs --follow.
  4. Press ctrl+C to exit.

If the containers are not running, use: dcf up to bring them online.

Once you are done debugging, press ctrl+C to exit and then run dcf down && dcf up -d to bring the containers back online in detached mode.

Django production logs​

If a bug happens in the backend code (i.e a 500 error), you can most likely find its traceback in the container's logs.

  1. Connect to the server using SSH.
  2. Run dcf logs backend --follow.
  3. Once you are done, press CTRL+C to exit.