Hello, small question. Sorry if I missed this in t...
# prefect-community
o
Hello, small question. Sorry if I missed this in the documentation, but is there a way to view what is in a work-queue? Im running an agent locally (2.7.0) with a --limit. Threading with context.
The work-queue became "unhealthy" per prefect UI, error logs showed that it was at capacity (hitting flow-run limit I set). However no runs were in the "running" state. The work queue stayed in this state for a long period of time. To debug this if it happens again, I was hoping to be able to see what the agent thinks it is currently churning on. Thanks!
m
Hey @Oliver E it sounds like you may have done this initially but you should be able to filter for running flows from the work-queues page in the UI, since that should ultimately be what's determining the limit from the work-queues perspective, the agent logic that handles the concurrency is I believe all on the work-queue side so I don't believe the agent is event aware of those limits afaik.
o
Thanks for the reply! I did not observe any running flows in the UI. Okay I think i might have had been smearing my terminology work-queue vs agent (im a newbie here) while I was debugging. However If i do see this again and there is no flow-runs running in the work-queue but my agent is not picking up new jobs, do you have any other recommendations on what I should take a peek at?
m
Hmm, worst case you can clear the concurrency limits from the work-queue and then recreate it, I'm pretty sure we're working on some visibility improvements around this, it's odd that you wouldn't be able to see any of the running flows, you might be able to query for running flow runs from the api as well to see if the UI is missing something for some reason. It definitely sounds like this could be a bug if you can put together a Minimum Reproducible Example it may be worth opening an issue on our github repo as well.
o
Just happened last night again agh. Thanks for the tips. Will open a github ticket if I can minimum repro. Thanks!