In Prefect 2 we are seeing flow runs that are stuck in a "Running" state, with all the tasks in that flow run marked as "Complete". The flow has a

timeout-seconds

defined that is already long past. Still the flow run stays in "Running" state. The k8s pod that was running the flow is already terminated. We saw this behaviour occasionally in Prefect 1 and were hopeful that it would improve in Prefect 2. Is there anything we should do to get this more reliable?

I'm wondering if there's actually a bug in the flow_timeout implementation...