Power spike while booting VM ?

[aharivel]

What happened?

On a OSP 18.0 installation with telemetry-power-metering activated (Kepler on the compute node), when we spawn a VM on the compute, we see a spike of power on the Node.

We know that the spike is wrong because 300K Watt would burn instantaneously the CPU and maybe the data center. But we are not sure about the last one.

Once the spike done the values seem correct.

Here the query, we run:

sum by (exported_instance) (rate(kepler_node_platform_joules_total{exported_instance="2b43fa7f966a"}[1m]))

What did you expect to happen?

A spike yes, but not a spike that high.

How can we reproduce it (as minimally and precisely as possible)?

• Install OSP 18
• Install the telemetry-power-monitoring operator
• spawn a VM

Anything else we need to know?

No response

Kepler image tag

Kepler v0.7.12

Kubernetes version

$ kubectl version
# paste output here

Cloud provider or bare metal

Bare metal

OS version

[cloud-admin@compute-0 ~]$ cat /etc/redhat-release Red Hat Enterprise Linux release 9.4 (Plow)

Install tools

Kepler deployment config

For on kubernetes:

$ KEPLER_NAMESPACE=kepler

# provide kepler configmap
$ kubectl get configmap kepler-cfm -n ${KEPLER_NAMESPACE}
# paste output here

# provide kepler deployment description
$ kubectl describe deployment kepler-exporter -n ${KEPLER_NAMESPACE}

For standalone:

put your Kepler command argument here

Container runtime (CRI) and version (if applicable)

Related plugins (CNI, CSI, ...) and versions (if applicable)

[sthaha]

May be related to #1774