Following on from the above, my flows are annotated with:

@flow(timeout_seconds=3600)

And yet I can see there are flows that have been in running for literally months. Im a bit confused as to how flows marked with a one hour timeout can be running for so long - has the way of declaring flows to timeout changed recently?? Now I dont have the workers available on the dask cluster to actually have hundreds of flows running, so I assume the tasks either finished/crashed/timed out, but prefect hasn't moved the actual state of the parent flow. Confused about how I can ensure this stops happening

Is it actually running or is it dead and the state just hasn't updated?

The ones shown here should be very dead and its just prefect's state which isnt updating