You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi. Livekit team. I hope this message finds you well.
I would like to report a behavior that I noticed, which may indicate a possible bug:
When there are no workers available, the jobs are enqueued, and as soon as workers become available, these jobs are assigned to them. However, I observed that jobs that do not have associated users are being assigned to workers. Is this the expected behavior?
Here are some session IDs that exhibited this behavior:
RM_PN67Yt9ePuKL
RM_ra4Hhth49Bod
RM_WKEcLVhppvjC
Here are the steps that I used in my test, if you want to reproduce this issue:
set the worker options to load_threshold=0.65 and the load_fnc to a function that returns the cpu usage
execute python agent-worker.py start
run as many playgrounds as possible til the worker is full. Don't close any of these.
run a last one playground. If the worker is full, you'll notice in the interface that the audio keeps loading (save the room name, you'll need it for the logs):
wait for a time (lets say 10-20 seconds) and close the last playground. Close the other ones.
In the session logs in the livekit cloud console, search the record with the room name that you saved earlier. You'll see something like this:
Thanks in advance for your attention, let me know if you need any additional details or clarification.
The text was updated successfully, but these errors were encountered:
Hi. Livekit team. I hope this message finds you well.
I would like to report a behavior that I noticed, which may indicate a possible bug:
When there are no workers available, the jobs are enqueued, and as soon as workers become available, these jobs are assigned to them. However, I observed that jobs that do not have associated users are being assigned to workers. Is this the expected behavior?
Here are some session IDs that exhibited this behavior:
Here are the steps that I used in my test, if you want to reproduce this issue:
load_threshold=0.65
and the load_fnc to a function that returns the cpu usagepython agent-worker.py start
Thanks in advance for your attention, let me know if you need any additional details or clarification.
The text was updated successfully, but these errors were encountered: