Collectd.log platform cpu usage value above 100%

[Bo Yuan Chang]

Brief Description

The platform CPU usage was indicated above 100% in collectedd.log, Should be normalized to a value in between 0 and 100 for better readability.

Severity

<Minor: System/Feature is usable with minor issue>

Steps to Reproduce

Run a task that computes heavy CPU power

Expected Behavior

Should show a number between 0-100

Actual Behavior

Entry exceeding 100 from time to time

Reproducibility

Reproducible

System Configuration

DC

Load info (eg: 2022-03-10_20-00-07)

22.06

Last Pass

N/A

Timestamp/Logs

info platform cpu usage plugin Usage: 98.8% (avg per cpu); cpus: X, Platform: 106.4% (Base: 97.6, k8s-system: 8.8), k8s-addon: 0.0

Alarms

Warning of memory host

Test Activity

N/A

Workaround

Manually normalize the number i.e. divide the log number by number of cpu cores.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/c/starlingx/monitoring/+/836388

[Ghada Khalil (gkhalil)]

Screening: opened as a minor issue, so will not hold up the stx.7.0 release