When I first set up Sentry with our app, I configured everything from the production instance to go to a single Project in Sentry. Django, celery, the browser JavaScript code, etc. After all, it’s all integrated; whether an error is reported from python or javascript, it might be part of the same user workflow.
But as we add a mobile app, other integrations, put different products on different release cycles, I wonder if we’re outgrowing the single-project organization.
To a large extent, I expect “separate project, or just a tag?” is a matter of taste. I also expect there are features where one way clearly works better than the other.
For example, if different products have different version numbers (release identifiers), is that a clear signal that they should be in different projects? Or will Sentry be okay with having multiple releases reporting to a single project at the same time? Seems like it would mess with its ability to decide a new release has happened.
But if we do split up things into multiple projects, will we lose visibility on correlations between browser and server-side errors, or if a single user is having similar errors among several of our products?
Thanks for any guidance you can give us on this.