Skip to main content

Kured

Kured (KUbernetes REboot Daemon) is an open-source Kubernetes

node reboot controller (infrastructure automation) that automates safe, coordinated restarts of cluster nodes when Operating System (OS) updates require a reboot.

  • Automates detection of pending reboots on Kubernetes nodes and triggers controlled node restarts (infrastructure automation).
  • Uses Kubernetes DaemonSet and leader election for coordinated node reboot orchestration (cluster operations).
  • Integrates with Kubernetes cordon, drain, and uncordon workflows to preserve workload availability during reboots (workload scheduling).
  • Supports configuration via Kubernetes resources and flags for maintenance windows, reboot strategies, and notifications (cluster management).
  • Operates within CNCF’s cloud native ecosystem as a utility for managing node lifecycle in Kubernetes environments (cloud native infrastructure).

More About Kured

Kured is an open-source Kubernetes node reboot daemon (infrastructure automation) that addresses the operational need to reboot cluster nodes in a controlled way when OS updates or kernel patches require a restart. It focuses on the node lifecycle portion of Kubernetes operations, where unmanaged reboots can affect application availability and cluster stability.

The project runs as a DaemonSet on Kubernetes (container orchestration), which ensures that an instance of Kured is present on each node in the cluster. Each instance monitors for signals that a reboot is required, such as the presence of a reboot flag file exposed by the underlying OS (node management). Through built-in leader election (cluster coordination), Kured designates a single active controller at a time to manage reboots, preventing simultaneous restarts that could reduce cluster capacity.

When a reboot is needed, Kured orchestrates a sequence that uses Kubernetes-native primitives. It cordons the node to prevent new pods from scheduling, drains the node to evict existing workloads where policies permit, initiates the node reboot, and then waits for the node to return before uncordoning it (workload scheduling and maintenance orchestration). This workflow aligns with Kubernetes operational practices and helps preserve application continuity.

Kured exposes configuration options for administrators to define behavior such as maintenance windows, reboot strategies, and optional notifications (cluster management). These settings can be supplied through Kubernetes manifests and command-line flags, allowing integration into GitOps pipelines and existing cluster management processes. The project is typically deployed in clusters that rely on automated OS patching where reboots are frequent and coordination is required.

Within enterprise and institutional environments, Kured is used as a utility component of platform engineering and Site Reliability Engineering (SRE) workflows (platform operations). It supports clusters running on various infrastructure providers compatible with Kubernetes, because it operates at the Kubernetes Control Plane (KCP) and node level rather than being tied to a specific cloud. As a project in the Cloud Native Computing Foundation (CNCF) ecosystem (cloud native infrastructure), it aligns with CNCF’s focus on cloud native architectures and standard Kubernetes practices.

From a directory and taxonomy perspective, Kured belongs in categories such as Kubernetes tooling, node lifecycle management, and infrastructure automation. It interacts with core Kubernetes APIs for nodes and pods, uses DaemonSets and leader election for distributed coordination, and focuses on safe, repeatable node reboot workflows that can be automated across clusters of various sizes.