Node State Metrics

Available in v1.30.8+

Node State is an observability sub-feature that exposes node-level health signals directly from the host: load average, runnable and blocked process counts, and Pressure Stall Information (PSI). Metrics are collected by the scaleops-network-monitor DaemonSet running on each node.

Node State runs independently — it does not require network monitoring, API observability, or workload identification to be enabled.

Installation

Node State requires the scaleops-network-monitor DaemonSet; enabling nodeState alone is sufficient to deploy it.

Option A — Node State only

If using a values file:

global:
  observability:
    nodeState:
      enabled: true

If using a Helm flag:

--set global.observability.nodeState.enabled=true

Option B — Full observability

Enables network monitoring, API monitoring, workload identification, and Node State together.

global:
  observability:
    enabled: true

Example upgrade command

helm repo update scaleops
helm show crds scaleops/scaleops | kubectl apply --force -f -
helm get values scaleops -n scaleops-system -oyaml | \
helm upgrade scaleops scaleops/scaleops \
  --set global.observability.nodeState.enabled=true \
  -n scaleops-system -f -

For additional configuration options (tolerations, node selectors, etc.), see the Helm Configuration Reference.