Florian Jetter

> @fjetter Is it? This seems very common on HPC. Could probably be caught by fixing this _once_ in dask-jobque. I doubt that there are many HPC users with a...

IIUC, this issue is about even more than just memory sampling, isn't it? theoretically we could sample all kinds of metrics and plot them

Haven't seen a failure in the past 30days. Closing

+1 for a dedicated `testing` module with a limited public API and the deprecation of `utils_test`

> Which is why I'd like this new API to use contextmanagers in place of fixtures I also support this. I don't consider exporting fixtures a "sane" API. It's way...

> I realised that _transition_from_executing is conceptually flawed. possible > Except that the assertion will fail in case of scatter. Why is scattering even transitioning via executing? That sounds wrong....

I believe this is a duplicate of https://github.com/dask/distributed/issues/4439 > The task is larger than 60% of memory_limit, so it's spilled immediately. > However, it fails to pickle back. I don't...

> I still need to figure out a reliable way to only track worker metrics when the workers are not idle. This would probably be a quite significant change to...

> > I'm hesitant to introduce full timeseries metrics for performance and memory usage reasons. > > Do these metrics qualify as timeseries metrics? I honestly didn't realize we were...

@charlesbluca sorry for the late follow up Yes, I believe you are right and my concerns were unfounded. Sorry for holding this up for so long. Can you rebase, please?