When you deploy a Kubernetes cluster, what you are deploying are nodes, which run the containerised applications.

Each cluster has two types of nodes:

1. Master node(s) host the Kubernetes Control Plane that controls and manages the whole Kubernetes system
2. Worker node(s) host the pods which are the components of the application.

graph TD
	subgraph WN[Worker Node]
		direction LR
		kubelet[kubelet] 
		runtime[Container Runtime] 
		proxy[kube-proxy]
	end
	
	subgraph MN[Master Node]
		direction LR
		api[kube-apiserver] 
		etcd[etcd] 
		scheduler[kube-scheduler] 
		controller[kube-controller-manager]
	end

Kubernetes Control Plane

The Control Plane controls the cluster and makes it function.
It has the following components:

1. kube-apiserver which you and the other Control Plane components communicate with
2. etcd a reliable distributed data store that persistently stores the cluster configuration.
3. kube-scheduler watches for newly created Pods with no assigned node, and selects a node for them to run on.
4. kube-controller-manager which performs cluster-level functions, such as replicating components, keeping track of worker nodes, handling node failures, and so on

Node Components

1. kubelet an agent, which runs on each node in the cluster and talks to the API server and manages containers on its node
2. kube-proxy which load-balances network traffic between application components
3. Container runtime : docker, rkt, etc.