-
Notifications
You must be signed in to change notification settings - Fork 23
Open
Description
We are using a scale set in an AKS cluster to run a system that runs untrusted code in different programming languages. We have had some incidents caused by the CPU usage reaches 100% but we don't know exactly which processes are causing this. This issue is linked Azure/AKS#4253.
I found that the following performance counters are available in VMSS monitoring logs:
- % Processor Time
- Pct User Time
- Used Memory
- Virtual Shared Memory
- Pct Privileged Time
Most the counters are self explanatory, but we have been trying to use Pct User Time and Pct Privileged Time to track the CPU time used per process, but the following things are not clear:
- What the counter values mean: Are they the total CPU time used by a process since it started? How can we use them to track the variation of CPU usage per process name?
- What is the difference between
Pct User TimeandPct Privileged Time? - Are there any docs on each counter?
Metadata
Metadata
Assignees
Labels
No labels