Addon: vcluster

Adding the vcluster addon will disable kubelet monitoring for vcluster deployments where the kubelet is not available.

On vclusters, the kubelet runs in the parent cluster and is not exposed to the vcluster. This addon prevents false alerts and failed scrape jobs by disabling the kubelet ServiceMonitor and filtering out kubelet-related alert rules.

Usage

To enable this addon, add it to your instance configuration:

parameters:
  prometheus:
    addons:
      - vcluster
    instances:
      my-vcluster:
        prometheus:
          enabled: true
        nodeExporter:
          enabled: false
        kubernetesControlPlane:
          enabled: true
The vcluster addon requires no additional configuration. Simply adding it to the addons list is sufficient.

What it disables

The addon disables the following kubelet-related monitoring:

  • ServiceMonitor: serviceMonitorKubelet - Prevents attempts to scrape kubelet metrics

  • Alert Groups:

    • kubernetes-system-kubelet - System-level kubelet and node alerts (KubeletDown, KubeletTooManyPods, KubeNodeNotReady, etc.)

    • kubelet.rules - Kubelet recording rules (metric aggregations that depend on kubelet metrics)

  • All rules using kubelet metrics: Any recording rule or alert that queries job="kubelet" is automatically filtered out, including:

    • Recording rules for container CPU and memory metrics from the kubelet’s cadvisor endpoint

    • Alerts that depend on kubelet-specific metrics (e.g., KubePersistentVolumeFillingUp, KubePersistentVolumeInodesFillingUp)

See also