we have a Dynatrace SaaS instance monitoring some of our local hosts.
Today at 8.40, the CPU on one of the hosts spiked from 20% (normal usage) to 70% (still is at 70%).
Dynatrace didn't signal anything in the "Problems" section. Why?
I'm attaching a graph of the CPU.
Furthermore, in the Settings->Anomaly detection->Infrastructure, the "Detect CPU saturation on host" setting is active. Maybe 70% is not a sufficient value to trigger an alarm?
Solved! Go to Solution.
As you suggested yourself, 70% is not enough to trigger a problem.
You can find the defaults in the documentation. For CPU Usage on a Host it is 95% in 3 of 5 one-minute intervals.
You can change this setting environment-wide, for a host group or a single host.
Possibly the conditions for an alert to be generated are not met. You can create your own alerts in which you can define a more sensitive threshold and under the conditions that you want.