|Synopsis||Kubernetes Dashboard showing incorrect CPU and Memory values|
|Affected Version(s)||1.6 and earlier|
|Updated on||02-11-2021 Initial Advisory|
In Kubernetes versions that are packaged with Konvoy 1.6 and earlier, there is a known issue where certain metrics like CPU Usage and Memory Usage are double-counted per container.
This results in incorrect metrics in the Kubernetes dashboard which can also cause some unpredictable Autoscaling behavior.
ResolutionFor the time being, we recommend ensuring that all Prometheus and Grafana addons are installed with your Konvoy cluster. The metrics in Grafana can be used as the correct source of information.
This issue is also going to be fixed in future versions of Konvoy, so this article will be updated when we have a specific fix version.