sentry sdk initialized on the client should be carried over to the workers

We don't support this for any task queue. I'm not sure if it's a good idea because I don't quite know how the implementation would work. We also don't transfer scopes from clients to workers.

We don't support this for any task queue. I'm not sure if it's a good idea because I don't quite know how the implementation would work. We also don't transfer scopes from clients to workers.

OK. Could than at least the integration add some context to the events emitted by the workers? (like computation ID or even which part of the computation graph is currently executed)

Sure. I don't want to discard the idea of carrying over scopes if it makes sense, but carrying over SDK config is probably not expected.

I'm open to all kinds of ideas, but they should be non-invasive in the sense that we shouldn't just add new arguments to a job just because we need them on the other machine.

This issue has gone three weeks without activity. In another week, I will close it.

But! If you comment or otherwise update it, I will reset the clock, and if you label it Status: Backlog or Status: In Progress, I will leave it alone ... forever!

"A weed is but an unloved flower." ― Ella Wheeler Wilcox 🥀

Currently we do not have plans to support this, Sorry.