Aurora metrics in PMM showing double CPU utilization when compared with AWS

I noticed 100% CPU in PMM for an Aurora node and checking the same node in AWS RDS showing effectively half the CPU utilization. It seems like that the total (node) CPU is added to the breakdown CPU utilization (user, system, wait, etc). Why is this? Can this be changed? It’s weird seeing 111% CPU :slight_smile:

Hi Ivan,

It’s fixed in the next PMM release 2.22.0
https://jira.percona.com/browse/PMM-8584

2 Likes

Excellent, thank you!

2 Likes

It looks like the CPU utilization reported in PMM is still double that of in AWS. I noticed that when enhanced monitoring is disabled, the CPU metric is displayed correctly, but all other instance metrics are N/A. When enhanced monitoring is enabled, the missing metrics are populated, but the CPU utilization is showing double.

1 Like