Node Exporter reports high CPU Pressure

[dmfrey]

Bug Report

Description

Pretty standard kube-prometheus-stack with node-exporter enabled. Node Exporter Full dashboard setup in Grafana.

Node Explorer is showing very high CPU Pressure

Here's my FluxCD HelmRelease to deploy kube-prometheus-stack for reference: https://github.com/dmfrey/home-gitops/blob/main/kubernetes/homelab-k8s-001/apps/monitor/kube-prometheus-stack/app/helmrelease.yaml

Logs

Environment

• Talos version: [talosctl version --nodes <problematic nodes>]
  Client:
  Tag: v1.9.1
  SHA: 348472f
  Built:
  Go version: go1.23.4
  OS/Arch: darwin/arm64
  Server:
  NODE: 192.168.30.33
  Tag: v1.9.1
  SHA: 348472f
  Built:
  Go version: go1.23.4
  OS/Arch: linux/amd64
  Enabled: RBAC
  NODE: 192.168.30.32
  Tag: v1.9.1
  SHA: 348472f
  Built:
  Go version: go1.23.4
  OS/Arch: linux/amd64
  Enabled: RBAC
  NODE: 192.168.30.31
  Tag: v1.9.1
  SHA: 348472f
  Built:
  Go version: go1.23.4
  OS/Arch: linux/amd64
  Enabled: RBAC
  NODE: 192.168.30.41
  Tag: v1.9.1
  SHA: 348472f
  Built:
  Go version: go1.23.4
  OS/Arch: linux/amd64
  Enabled: RBAC
  NODE: 192.168.30.42
  Tag: v1.9.1
  SHA: 348472f
  Built:
  Go version: go1.23.4
  OS/Arch: linux/amd64
  Enabled: RBAC

• Kubernetes version: [kubectl version --short] 1.32.0
  Client Version: v1.32.0
  Kustomize Version: v5.5.0
  Server Version: v1.32.0

• Platform:

[smira]

How can that be a Talos Linux issue?

[dmfrey]

I was asked to move the Slack conversation to a GitHub Issue.

After some more digging, there is an existing Linux kernel issue regarding this. I believe a fix is in, but it hasn't yet made its way into the released kernel.

torvalds/linux@3f02039

[rothgar]

Do you know what kernel(s) this effects?