Is the baseline as simple as what it appears? Ie: 7 day recalc by default. So, If I'm sampling every 5 mins, this would be the average value of 12*24*7, or 2,016 samples? If so, using such a baseline for alerts would not be very helpful. One certainly can't determine normal/abnormal behavior if you roll everything up into a 7 day summary.
What I am hoping for is something more like an hourly baseline over 7 days. Ie: During the 3am-4am block, my baseline would be specific to that block: 12 samples * 7 days = 84 in my data set. In a perfect world, Every hour of every day for the last month (or two weeks, whatever) would have it's own baseline value. Then I could distinguish between work days AND work hours, nightly backups and updates, etc. In other words, what normal for 10am on a Monday is totally different than 10am on a Saturday.
Happy to open a ticket if no one has the details on this. I may be able to accomplish my end goal, or close to it, using custom alerts and triggers. In the end, just don't want callouts for "normal", and sometimes 99% for 4 hours IS 100% normal.
TIA
Gary
