We're using Raven & Sentry to track errors from React app And each deploy produces new js bundle:
- my_bundle.e1277d4630b7d8cbb771.js
- my_bundle.844d5be4607494ae004f.js
- ...
The same bundle can be served from different URL:
- https://my_website.com/en/my_bundle.e1277d4630b7d8cbb771.js
- https://my_website.com/fr/my_bundle.e1277d4630b7d8cbb771.js
- ...
As a result when we have some error on js, with the same stack trace but from different bundles (releases) or from the same bundle but served from different URLs /en/my_bundle.e1277d4630b7d8cbb771.js | fr/my_bundle.e1277d4630b7d8cbb771.js Sentry group them in a random way: errors are grouped in some batches: so from 100 events of the same error, we have ~ 40 different issues in sentry. For example, issues could be grouped in such a way:
First group:
- /en/my_bundle.e1277d4630b7d8cbb771.js
- /fr/my_bundle.e1277d4630b7d8cbb771.js
- /de/my_bundle.844d5be4607494ae004f.js
Second group:
- /de/my_bundle.e1277d4630b7d8cbb771.js
- /fr/my_bundle.844d5be4607494ae004f.js
Is there a way to configure Sentry to group errors in the more intelligent way, to group error with same stack trace as one issue?
I've found an approach to normalize file name in Raven before sending data in Sentry:
https://gist.github.com/danharper/a89685cc1028759345b8
But it seems a bit overcomplicated, cause as I see Sentry trying to group error from different files by itself.
