The Process Monitors for components have out of the box thresholds for:
- CPU
- Physical Memory
- Virtual Memory
- I/O Read
- I/O Write
- I/O Total
But one we're missing is for process count. This is an effort to implement a workaround to alert when a count is over a certain threshold.
Type: Custom SWQL
This is the basis of the alert in SWQL from the alert (but there are additional notes in the alert)
SELECT TOP 1000 [Component].Uri
, [Component].DisplayName
FROM Orion.APM.Component AS [Component]
WHERE [Component].ProcessEvidenceChart.AvgInstanceCount BETWEEN 1 AND 150
AND [Component].Application.Node.Vendor = 'Linux'
This can easily be tuned for different vendors (by changing the filter) or filtering for specific component names.
Usage
Component Monitors require Server & Application Monitor (SAM) or Hybrid Cloud Observability.
Out of the box thresholds for SAM Process monitors
