Arroyo creates slightly confusing stacktraces for application level errors

[JoshFerge]

https://sentry.sentry.io/issues/4252310678/events/ab963ef95d334ca1b8b20079b4421d41/?project=1

this error seems like it comes from the arroyo library code, even though it happened in our application level code. It would be nice if we could have more straighforward stacktraces from our consumer errors.

[untitaker]

[untitaker]

I think this is very similar to the grouping issues we had in native, with random signal handlers interrupting the work of the main thread. We should try and figure out some additional updates to the grouping algorithm to make those stacktraces group better (truncate a few threadbases) + also add logic to arroyo so that teh stacktrace becomes more readable