Description of problem:
The status controller of CCO reconciles 500+ times/h on average on a resting 6-node mint-mode OCP cluster on AWS.
Steps to Reproduce:
1. Install a 6-node mint-mode OCP cluster on AWS 2. Do nothing with it and wait for a couple of hours 3. Plot the following metric in the metrics dashboard of OCP console: rate(controller_runtime_reconcile_total{controller="status"}[1h]) * 3600
Actual results:
500+ reconciles/h on a resting cluster
Expected results:
12-50 reconciles/h on a resting cluster Note: the reconcile() function always requeues after 5min so the theoretical minimum is 12 reconciles/h
- blocks
-
OCPBUGS-29155 Redundant reconciles by CCO's status controller
- Closed
- is cloned by
-
OCPBUGS-29155 Redundant reconciles by CCO's status controller
- Closed
-
OCPBUGS-29156 Redundant reconciles by CCO's status controller
- Closed
- links to
-
RHEA-2024:0041 OpenShift Container Platform 4.16.z bug fix update