Closed

5 months ago

This isn't a feature request. It's a configuration question. Depending on the product in question (I'm guessing either Network Performance Monitor or Hybrid Cloud Observability from the screenshot), you should ask the question in that particular forum.

Closed as inaccurate type and in wrong Feature Request space.

High CPU usage not resulting in a critical node status

We don’t want the status of a Linux node (physical and virtual) to change to critical if there is a high CPU usage, as for some systems this is accepted behaviour.

Some things we've tried without the desired result :

Disabling Critical CPU Status in de Node Details.
Node Deails => Edit Node => Node Thresholds => CPU Load => Unselect Critical Status.

Disabling Node Status Contributors is not an option since some teams, like the Windows and Netwerk team, want the node status to be critical if there is a high CPU usage.

Also did some testing with "Virtualization Settings" and "Virtualization Global Thresholds".

smukkie

KMSigma 5 months ago

As an aside, I've got a Linux machine where I can play with these settings. Mine also runs "hot" because it's hosting Docker containers.

This is the before:

As you can see, it was triggering alerts because it was in a Warning status.

I configured your settings above on that Linux machine in my lab.

After the next polling cycle (9 minutes), it flipped from Critical to Up (Green) after the next polling cycle.

My lab is currently running SolarWinds Platform version 2023.1.

I also don't use any of the OOTB alerts in production environments. That way I can tune them as needed.

If you aren't seeing this activity, then this feels like a configuration problem or a bug. If you are still experiencing this issue after changing the settings and allowing the environment to poll the device and all child elements again, I'd contact support.
- Cancel
- Up 0 Down
- More
- Cancel