I have servers that have nightly maintenance windows, during which time I have a job that runs to defragment indexes and update statistics. There are consequences to running such maintenance jobs, such as PLE dropping to near zero, although it immediately climbs back up afterward. Another example would be that our SAN has some kind of nightly maintenance that causes latency to jump to a very high value for a minute or two at the same time every night.
How can I set up specific alerts, such as "All Metrics in a Category - Memory", specifying an exclusion window so I don't get alerted for values that would normally trigger it, but are completely normal during that window?