Description
GKE node NotReady troubleshooting is an agent skill for finding out why one or more Google Kubernetes Engine nodes report NotReady or Ready Unknown status. A NotReady node means its kubelet is not reporting correctly to the control plane, so Kubernetes stops placing new pods on it, which can cut application capacity and cause downtime. The skill works through a fixed diagnostic workflow. It inspects node conditions, events, kubelet and containerd logs and node metrics, then proposes a safe remediation as a kubectl or gcloud command, or as a change to a manifest for a person to apply. It is deliberately read-only. It never changes the cluster, drains, deletes or recreates nodes on its own. The skill first rules out an expected NotReady, such as a node that is newly provisioning, upgrading, being repaired, cordoned or scaling down, since these report NotReady briefly, and it treats the status as a fault only if it persists beyond the expected window. It is useful for on-call engineers and platform teams who need a structured first pass before escalating. It is published by Google on skills.sh under the google skills repository, with about 2.3K installs. The skills.sh listing shows Gen Agent Trust Hub, Socket and Snyk all passing.…
Highlights
- Inspects node conditions, events and logs
- Read-only diagnostics by design
- Proposes fixes for a person to apply
- Rules out expected NotReady states first
- Covers kubelet, containerd and node metrics
Ratings & Reviews
No ratings yet
Be the first to rate gke-node-notready.
People also visit
gke-cluster-autoscaler
Trigger on mention of GKE cluster autoscaler, node autoscaling
google-cloud-solution-guided-gke-ai-migration
Guides the migration of existing AI workloads
agent-platform-troubleshooting
Troubleshoots Google Cloud Gemini Enterprise Agent Platform issues
gke-node-notready FAQ
What does NotReady mean?
The node's kubelet is not reporting correctly to the control plane, so Kubernetes stops scheduling new pods on it.
Does the skill fix the node?
No. It is read-only and proposes a command or manifest change for a person to apply. It never drains, deletes or recreates nodes automatically.
Is every NotReady a fault?
No. Nodes that are provisioning, upgrading, being repaired, cordoned or scaling down can report NotReady briefly.
What does it look at?
Node conditions, events, kubelet and containerd logs, and node metrics.
How do I install it?
Run npx skills add https://github.com/google/skills --skill gke-node-notready.
What are the best gke-node-notready alternatives?
Popular alternatives to gke-node-notready include gke-workload-troubleshooting, gke-backup-dr and gke-cluster-autoscaler. See all gke-node-notready alternatives