Server monitoring
Kubernetes node monitoring with one DaemonSet
Deploy the agent as a DaemonSet and every node — including ones the autoscaler adds later — reports in automatically. Pair it with HTTP monitors on your ingress for an outside-in view of what users actually see.
What to monitor on Kubernetes
- 01Node CPU and loadDetect nodes that are overcommitted beyond their requests.
- 02Node memoryHigh memory triggers kubelet evictions before you notice.
- 03Node diskDiskPressure taints stop new pods being scheduled.
- 04NetworkPer-interface throughput including CNI interfaces.
Set it up
Create the secret
Store your API key in the cluster.
kubectl create secret generic statance --from-literal=key=YOUR_API_KEYApply the DaemonSet
Creates a statance namespace-scoped DaemonSet with read-only host mounts and tolerations for all nodes.
kubectl apply -f https://statance.com/agent/kubernetes.yamlWatch nodes appear
Each node registers with its hostname.
kubectl get pods -l app=statance-agent -o wide
Recommended alerts for Kubernetes
- Server stopped reporting for 5 minutes (down)
- CPU above 90% for 10 minutes
- Memory used above 90% for 5 minutes
- Any disk above 85% used
- Node stopped reporting (drained, crashed or scaled in)
Send alerts to
Kubernetes monitoring FAQ
What happens when the autoscaler removes a node?
The server stops reporting and is marked down. Delete it from the dashboard or via the API, or simply leave down-alerts off for autoscaled pools.
Does it work on managed clusters (EKS, GKE, AKS, DOKS)?
Yes, on any cluster that allows hostPath mounts and hostPID. GKE Autopilot does not allow these.
Know before your users do.
Free forever for 3 servers and 10 monitors. No credit card. Install the agent in under a minute.