I am curious if anybody has taken the time to write an alert that looks at load average on a Linux/Unix server that then takes the load value and divides it by the number of CPU's in the box?
I have the same requirement. Need to configure alerting when the load average exceeds the number of processors (or when the normalized load averages is greater than 1).
This is an old question, but I have written an alert for this.
I want to alert on:
Custom SWQL Alert
Set up your SWQL condition:
SELECT Nodes.Uri, Nodes.DisplayName FROM Orion.Nodes AS Nodes
WHERE Nodes.CustomProperties.Environment = 'Linux Non-Prod System'
AND Nodes.LoadAverage1 > (Nodes.CPUCount * 6)
OR Nodes.CustomProperties.Environment = 'Linux Non-Prod System'
AND Nodes.LoadAverage5 > (Nodes.CPUCount * 4)
AND Nodes.LoadAverage15 > (Nodes.CPUCount * 3)
It works great for us. Let me know if that helps.
We have a custom property Environment which we have populated with Linux Non-Prod or Linux Prod. I have two of these set up, one for non production and one for production.