GKE Node NotReady skill: what it does and how to install it

What it does. Checks whether a node's status is an expected transient, then guides read-only investigation of node conditions, events, kubelet and serial-console logs, and resource metrics. Maps evidence to likely causes and proposes a remediation for a human to apply, with links to supporting logs and metrics.

When to use it. For GKE nodes that remain NotReady or Unknown, stop reporting status, or cause workloads to be evicted or stuck Pending. Not for pod-level application failures, autoscaler decisions, or non-GKE compute.