Intermittent timeouts on requests

I have a custom docker image serverless endpoint. I am sending a payload with the python package endpoint.run_sync(payload, timeout=60). I currently have 0 active workers. I can typically send the first request. After completion, if I send a following request before that worker times out, it will often timeout without ever logging that the main function started. Basically, it seems like the message never gets in the RunPod Queue. What could be causing this behavior? How can I avoid it (or debug it)?
Logs are attached - this case is 2 successful requests, then a third request just times out - it seems like the request never gets to the queue (no logs).
No description
4 Replies
nerdylive
nerdylive3mo ago
Any endpoint is? Id Have you resolved this yet?
Kid Smart AI
Kid Smart AIOP3mo ago
I changed my process, i just repeatably request with 5 second timeout.
nerdylive
nerdylive3mo ago
So with longer timeouts it works?
flash-singh
flash-singh3mo ago
runsync max timeout is 90s, we plan to increase this to 3mins in near future

Did you find this page helpful?