Ahmed Ezzat

Results 44 comments of Ahmed Ezzat

Seems like they're timeout issues not something failing

Looks like I didn't provide much information, this issue is about `scheduler worker` communication, I'd like my cluster to communicate over `ws` rather than the hardcoded `tcp` protocol since you...

I'm fully aware that https://distributed.dask.org/en/latest/resources.html exists however it doesn't trigger autoscaling.

Looks good to me, we should also subtract pending workers from the number of workers passed to `retire_workers`