We want to alert on average CPU usage > some percentage, but only if that's sustained over more than ## seconds. We have a lot of cases where the CPU spikes above some arbitrary number, but then dips down shortly afterwards. Those one-off spikes aren't necessarily a problem, but we don't want to be alerted on them unless that spike continues for more than a couple of seconds (or some other interval we'd define).
What's the best way to handle that in Applications Manager? It seems like there should be a way to catch that.
Also, is there a way to catch that for both "total CPU" and "individual cores"? We'd set them up slightly differently, but if one core is pegged for some time that could be a cause for an alert whereas if _all_ cores are above that threshold we'd want to know sooner.