Error Tracking

In order to track errors in applications we rely on Sentry as a central hub for automatically collecting exception raises or other problems. From here, we use the build-in GitLab integration to create issues in GitLab for problems to be fixed.

We use Sentry's Slack integration to push exceptions to Slack so that all relevant developers can see them. Either they are pushed into the #dev_$project or the #dev_$team channels.

All of our developers are able to access Sentry for their projects and should take it upon themselves to fix bugs that occur as soon as possible.

Rules for Error Tracking

1. If it runs in production, it must catch errors and report them to Sentry.
2. If you have extra information (e.g. user context, request specific things) report the context into Sentry along with all errors. See: Sentry Docs on Context
3. Do not ignore errors, even if you think they are harmless.
4. Make sure to setup Alert Rules to the Slack Integration (see image below).

Sensible rule for most projects

1. Setup notification to be send each time an event is triggered, not only first time (which is default).
2. Push to the correct channel. Make sure to /invite @Sentry into that channel.
3. Set reasonable period to avoid spamming channels when an error occurs constantly.