OSS Alternative - Discover Top Open Source Alternatives to Popular Software

kubernetes/node-problem-detector

A Kubernetes daemon that detects and reports various node problems to the apiserver, making node health visible for improved cluster management.

Core Features

Detects a wide range of node issues including hardware, kernel, and container runtime problems.

Reports persistent node issues as `NodeCondition` and transient issues as `Event` to the Kubernetes apiserver.

Utilizes modular problem daemons like SystemLogMonitor and SystemStatsMonitor for specific problem detection.

Can run as a Kubernetes DaemonSet or standalone on each node.

Integrates seamlessly with Kubernetes cluster management for proactive issue handling.

Detailed Introduction

Kubernetes nodes can suffer from various underlying issues such as hardware failures, kernel deadlocks, or unresponsive container runtimes, which often remain invisible to the cluster's scheduling layers. Node-problem-detector addresses this by running as a daemon on each node, actively monitoring for these problems. It then reports critical, persistent issues as `NodeCondition` and temporary, informative ones as `Event` directly to the Kubernetes apiserver. This visibility empowers upstream Kubernetes components to make informed decisions, preventing pods from being scheduled onto unhealthy nodes and enabling more robust cluster operations.