Hey guys, anyone experiencing some runs that get stuck in in_progress for an absurd amount of amount time? It’s usually when I ask my assistant something sort of unrelated but trivial.
I am using 2 functions as well as 2 uploaded files to gpt-4-1106-preview.
What i’ve done for now is set a hard timeout of around 30s and then cancelled the run, reported back to the user that it couldn’t answer the question but just wondering if this is something i’m doing wrong? I’ll check back quite a bit later and the runs are still in_progress or stuck on cancelling.
Trying it out in the playground it actually just gets stuck on Get run's steps in the logs with no movements.
I am using a polling mechanism in my API to check the status like others have done. Any tips/advice to make this more robust? Is it something i need to tell my assistant to do?