Problem:

Prometheus VM Metrics are not accurate whatsoever where cpu usage metrics at a percentage level are exceeding 200000% within the Harvester UI VM metrics.

Solution:

fix incorrect cpu usage unit, upgrade path fix is here: harvester/harvester#7250

Related Issue:
harvester/harvester#7086

Test plan:

• Prepare a harvester cluster with this patch
• Enable rancher-monitoring addon
• Prepare some VMs (e.g. 3 VM)
• Goto Grafana UI and wait for a couple of minutes
• Check the value of cpu usage should fall in range [0, 100] in both Harvester VM Dashboard and Harvester VM Info Detail dashboards (during vm startup, the cpu usage may slightly go over 100%, after a while it should dramatically drop to a low value, like 2%)

---

This is an automatic backport of pull request #908 done by [Mergify](https://mergify.com).