Kubelet

The kubelet is the primary node agent that runs on every worker node in a Kubernetes cluster. Its job is to ensure that the containers described in the pods assigned to its node are running and healthy.

How It Works

The kubelet registers its node with the control plane and continuously watches the API server for pods scheduled to that node. For each assigned pod, it instructs the container runtime (through the Container Runtime Interface, CRI) to pull images and start, stop, or restart containers as needed.

The kubelet also runs the pod's health probes, liveness, readiness, and startup checks, and reports node and pod status back to the control plane. It manages pod resources, mounts volumes and Secrets, and reports node conditions such as memory or disk pressure. Notably, the kubelet manages only pods created through Kubernetes; it does not manage arbitrary containers started outside its control.

Why It Matters

The kubelet is the bridge between the cluster's declarative desired state and the concrete containers running on a machine. Without it, the control plane's scheduling decisions would never be carried out. Its health and accurate status reporting are essential for scheduling, self-healing, and autoscaling to function correctly.

Because it runs on every node and has privileged access to the host, the kubelet is also an important security boundary that must be authenticated and authorized carefully.

Related Terms

The kubelet runs pods on its node by driving the container runtime, takes direction from the control plane, and works alongside per-node agents deployed via a DaemonSet.