Kured is a Kubernetes daemonset that automates safe node reboots based on OS-level reboot requirements to maintain cluster stability and security.
Kubernetes Reboot Daemon
Kured is used by Kubernetes cluster operators to automate the reboot of nodes when the underlying operating system signals a reboot is required, such as after security patches or updates. It ensures reboots happen safely by cordoning and draining nodes, coordinating reboots one at a time, and optionally deferring reboots based on active alerts or running pods, thus minimizing disruption in production environments.
Kured requires appropriate permissions to cordon, drain, and uncordon nodes in the Kubernetes cluster. It is recommended to integrate with Prometheus for alert-based reboot deferral to avoid disruptions during critical incidents. Since it reboots nodes, careful configuration and testing in staging environments are advised before production deployment.
Visit https://kured.dev/docs/installation/ for detailed installation instructions
Deploy kured as a Kubernetes daemonset in your cluster
Configure kured according to your environment and reboot policies
Ensure your nodes have the reboot sentinel file or command configured
Optionally configure Prometheus alert integration for reboot deferral