I have seen this occur randomly in the past but have never found a root cause. Has anyone had the properties for "IsLevel1State" or "IsLevel2State" get stuck in a true state? Traditionally a reboot of our main polling engine fixes the issue. I'm curious though if anyone else has ran into this.
Basically a server had a momentary CPU spike, which happened to get caught by polling. So the two state values going to true makes sense. They should have moved back to False on the next poll. Since the flag was true, it then eventually tripped our CPU based alert since the value was true for more than 30 minutes.
I've already gathered the agent logs, screen shots, and diagnostic logs and am uploading them to a support case. I figured it wouldn't be a bad idea to throw this out to thwack as well and see if any of you have thoughts on the matter.
What a healthy node looks like:
