Skip to content

Latest commit

 

History

History
46 lines (33 loc) · 2.02 KB

README.md

File metadata and controls

46 lines (33 loc) · 2.02 KB
Raw

Architecture

A Kubernetes cluster consists of a control plane and a set of worker machines (commonly referred as nodes).

The control plane manages the worker nodes and pods. Its components make decisions about the cluster as well as detecting and responding to cluster events.

Nodes run containerized applications. Every cluster has at least one worker node.

kubernetes architecture

Components

  • API server - Works like a "Cluster gateway". Interact with clients (kubectl, UI, etc.) and act like a auth gatekeeper.

  • Cloud controller manager - Lets you link your cluster into your cloud provider's API, and separates out the components that interact with that cloud platform from components that only interact with your cluster. This component is optional.

  • Controller manager - Detect cluster state changes. Watch the shared state through the API server and make changes attempting to move the current state to the desired state.

  • etcd (the “cluster brain”) - A key value store. Stores every cluster change (when a pod dies, when a new pod is scheduled, etc.) Other control plane resources depends on the data store here to take an action.

  • Kubelet - Interacts with containers and node. Responsible for running pods, allocating container resources (CPU, memory, etc.) and starting the pod with a container inside.

  • Kube proxy - Forwards request from services to pods. Has an intelligent forwarding logic to balance pod calls between worker nodes.

  • Scheduler - Watches for newly created Pods with no assigned node and selects a node for them to run on. Scheduler just decides on which node the new pod will be placed and then send a request to Kubelet (that actually does the pod schedule).

  • Container runtime Engine that run the containers on the hosting operating system. The most used container runtimes are docker, containerd and cri-o.