We use Sentry’s Python SDK and Flask/Celery integrations, which by default send all exceptions. This includes 400 code exceptions like BadRequest and NotFound.
To the best of my understanding, the errors in Sentry should be actionable, meaning that you can resolve them. But this is not the case with such exceptions, because (in ideal world) they are user errors.
We were about to filter those errors, but then we had a production issue resulting in a BadRequest exception. Having all the data Sentry includes with each event was a god send and allowed to resolve this issue in minutes.
I wonder what’s the best practice here? Is there a place for non actionable events/errors in Sentry?