robusta icon indicating copy to clipboard operation
robusta copied to clipboard

Published 20 hours ago • verified publisher icon robusta-dev

Bug in HPA alert enrichment

Open luigiaparicio opened this issue 2 years ago • 0 comments

Describe the bug I have a Deployment and HPA configured to autoscaling Pods by CPU utilization.

When HPA scale the Deployment to the maximum number of replicas, Robusta notifies "HPA reached max replicas" and inside the message nforms a "Current avg cpu utilization" percent that is lower than the HPA target CPU configured in HPA. In my case target is configured at 80%, and Robusta alerts say 18% current

To Reproduce Steps to reproduce the behavior:

  1. Generate load inside the pod, to increase the CPU usage
  2. HPA scales up to max replicas because current is higher than target.
  3. Wait for Robusta to detect this behavior and alert
  4. Look in the alert message and see that current is lower than HPA target cpu threshold

Expected behavior Current percent informed should be higher than HPA target cpu threshold

Screenshots Screenshot 2023-02-10 at 10 12 11

Additional context This tests were done in an OpenShift cluster version 4.10 and notifications received in Slack

luigiaparicio avatar Feb 10 '23 15:02 luigiaparicio