Hi 👋 We have a critical issue: https://github.com/PrefectHQ/prefect/issues/12716 Out of nowhere runs are failing when trying to use an imported python module (any module). Any ideas what might be causing this?

I'm not 100% sure what's causing this, but I do have a top candidate I suspect might be it. Would it be possible to make a new work queue not named

default

in the work pool where this is happening and move your deployments over to it, or restart your agent but include a work pool name in the startup command? I've seen some issues lately with intermittent failures in workspaces that have both agents and worker-style work pools, and both of those pools have queues named

default

@Kevin Grismore Regarding the “agent” advice, we are using ECS push work pool, so there is no agent involved right?

there is still an agent polling queues in your workspace though, right?

I am not sure. I think the only thing we have is a “ecs cluster” we’ve created in our aws account. Prefect just sends new job definitions to it.

hm, let me follow up in dms