Sentry init only in __init__.py suffices?

SDKs
1

Quite new to sentry and could not find my question in the docs.

Let’s say we have a project with 30 python scripts. Does it suffice to put sentry_sdk.init(..) in your __init__.py and will it track the errors in all the 30 python scripts? Or do we have to put sentry_sdk.init in every script to track the errors?

Furthermore, I am using the @serverless_function since we run our projects on Azure functions.

2

You need to initialize the Python SDK once per OS process, and there ideally as early as possible in the import-chain.

So you can put init() into just one file, but every process needs to import that file at some point.

I don’t know your project layout but __init__ is usually imported all the time, so it sounds about right to only put it there.

3

Thanks for the clarification, will try and share the results here.

Would be nice if there was an example or at least something mentioned about this in the docs. I think this is a vital point of using Sentry in bigger projects.