# Node-problem-detector

Node-problem-detector is a Kubernetes add-on daemon that runs on each node to detect node problems and report them to the API server, making infrastructure issues visible to the cluster management stack. It is open source and self-hosted, installable on Kubernetes (Helm).

Canonical page: https://ourlocalhost.com/app/4737/node-problem-detector

## Description

Node Problem Detector is a Kubernetes add-on daemon that runs on each node to detect node problems and report them to the API server, making infrastructure issues visible to the cluster management stack. It helps operators spot failing nodes early, before they disrupt the workloads running on them.

## Platform availability

Available on 1 of 10 platforms: Helm.

## Health score

77/100 (Good), based on 4 of 5 factors.
- Maintained: 100 (last commit 19 days ago)
- Popular: 71 (3,483 GitHub stars)
- Easy to install: 25 (1 install platform, install notes)
- Security check: 89 (no serious findings, CI, tests, OpenSSF Scorecard 7.2/10)

## Links

- GitHub: https://github.com/kubernetes/node-problem-detector
- Kubernetes (Helm): https://artifacthub.io/packages/helm/deliveryhero/node-problem-detector

## Common questions

### Can node-problem-detector be deployed without a DaemonSet?

It can run as a DaemonSet or standalone.

### What does node-problem-detector watch on each node?

It watches kernel logs, systemd units, kubelet health, and user scripts.

### Can it detect unhealthy GPU nodes?

It has been discussed with related tooling for detecting unhealthy GPU nodes and exploring automatic remediation.

### Where do I look for node health information after setup?

You can observe node and pod status in the Kubernetes cluster.

---
Source: Our Local Host (https://ourlocalhost.com/app/4737/node-problem-detector), the self-hosted and open-source app directory.
