Kubelet

kubelet process in Kubernetes is responsible for running containers on the node

• Configuration: /etc/kubernetes/kubelet.conf
• Logs:
  • /var/log/kubelet.log
  • GKE: gcloud logging read projects/your-project/logs/kubelet
• Port: kubelet listen on port 10250

• Installation Ubuntu: snap install kubelet --classic (snap stop kubelet.daemon or snap start kubelet.daemon)
• Kubelet logs

Commands

systemctl start snap.kubelet.daemon.service
systemctl stop snap.kubelet.daemon.service

snap start kubelet.daemon
Started.
systemctl status snap.kubelet.daemon.service

Networking

The Container Network Interface (CNI) plugin is selected by passing kubelet the --network-plugin=cni

Related terms

• Liveness probe: (livenessProbe)
• Readiness probe: (readinessProbe)
• systemctl stop snap.kubelet.daemon.service
• kubectl describe pods | grep Unhealthy
• Kubelet Container Runtime Interface (CRI)
• systemctl status kubelet
• gcloud logging read projects/yourproject/logs/kubelet
• Kubernetes controller manager
• Conditions: kubelet is posting ready status. AppArmor enabled
• --kubelet-insecure-tls
• MemoryPressure
• kubectl get events Starting kubelet
• cloud-node-lifecycle-controller
• Node-pressure Eviction
• kind: KubeletConfiguration
• oom_score_adj
• Kubelet stopped posting node status.

Activities

• CKA v1.15: Understand LivenessProbes and ReadinessProbes

News

• 05/2022 Kubernetes 1.24: dockershim removed from kubelet

See also

• kubelet, Pod Lifecycle Event Generator (PLEG), Kubelet conditions, /etc/kubernetes/kubelet.conf, /var/log/kubelet.log, Kubelet Container Runtime Interface (CRI), crictl, context deadline exceeded, E1122, E1124, Starting kubelet, NodeHasDiskPressure, NodeHasInsufficientMemory, rebooted, Node-pressure Eviction, Kubelet logs, manager: kubelet
• Kubernetes Components, kube-proxy, kube-scheduler, kube-apiserver, kubelet, kubectl, kubeadm, Kubernetes controller manager, Kubernetes controllers, Kubernetes Addons, Kubernetes Topology Manager