I have a HDP 3.1 cluster and it seems that the fair policy isn't behaving as expected or YARN is misconfigured, since some users/applications/jobs are consuming more resources than we supposed it to use. So how do I debug/monitor YARN in a way that when some behaviour occurs I get notified, as well as, log the information I need to debug the queues so I can check if there is, in fact, a problem?
I appreciate any help.
