Control Plane vs. Worker Nodes: Responsibilities and Component Breakdown

Quick Answer: A Node is a physical or virtual machine in the cluster. Master nodes (control plane nodes) run cluster management components; worker nodes run your application pods.

Detailed Answer:

In Kubernetes, nodes are the machines (EC2 instances, VMs, or bare metal) that form the cluster. Every node runs the kubelet (the node agent), kube-proxy (network routing), and a container runtime (containerd). The distinction between master and worker is about which additional components run on the node:

• Control plane nodes additionally run: kube-apiserver, etcd, kube-scheduler, kube-controller-manager. In managed services like EKS, these are fully managed by AWS — you never see these nodes.
• Worker nodes run your application pods. On EKS, these are the EC2 instances in your node groups that you pay for and manage.

On EKS Specifically (Your Context):

The control plane is AWS-managed — you don't provision, patch, or back up control plane nodes. You only manage worker nodes (via managed node groups or Karpenter). This is one of EKS's biggest operational advantages.

kubectl get nodes                      # List all nodes
kubectl get nodes -o wide              # With IP, OS, container runtime
kubectl describe node <node-name>      # Full detail: conditions, capacity, allocated pods
kubectl cordon <node-name>             # Mark node unschedulable (drain prep)
kubectl drain <node-name> --ignore-daemonsets --delete-emptydir-data
kubectl uncordon <node-name>           # Re-enable scheduling after maintenance

Key Takeaway: In managed Kubernetes (EKS, GKE, AKS), you only manage worker nodes — the control plane is the cloud provider's responsibility.