We are using AIOPS Memory metric for our Support portal application pods but we've found some error with Memory % (rounded) metric. Below are the details.
As per this KB : https://techdocs.broadcom.com/us/en/ca-enterprise-software/it-operations-management/application-performance-management/10-7/implementing-agents/infrastructure-agent/kubernetes-monitoring/kubernetes-monitoring-metrics.html
Memory % = Memory usage divided by Memory limit.
But if you see attachment, Our application container's Memory usage is around 2.7 GB and Memory limit is around 6 GB.
So, if we divide as per the KB , Memory % (2.7 divided by 6) should be 45%.
But in the AIOPS console, we are seeing 100% which we believe is a false metric.
Release : 22.1
The below wiki should provide more information about the UMA memory metrics: