Lecture Notes: Kubernetes Architecture

Introduction

• New chapter: Understanding Kubernetes from scratch, focusing on architecture and its role in DevOps.
• Kubernetes developed by Google.
• Emphasis on understanding architecture to grasp Kubernetes' orchestration capabilities.

Kubernetes Architecture Overview

• Control Plane ("Brain" of Kubernetes):
  • Central control component in Kubernetes.
• Worker Nodes:
  • Execution environment where applications run.
• Example: A single node cluster (Minikube), can be expanded to multi-node cluster.

Key Components of Control Plane

1. Kube-API Server

   • Front end of Kubernetes brain.
   • Manages communications and scheduling with nodes.
   • Interfaces with other components: Controller Manager, etcd, Scheduler.

2. etcd

   • Backend database storing data as key-value pairs.

3. Scheduler

   • Assigns newly created pods to nodes.
   • Ensures pods are scheduled on appropriate nodes.

4. Controller Manager

   • Distributes load away from Kube-API Server.
   • Contains several controllers:
     • Node Controller: Detects node failures, manages node status.
     • Job Controller: Manages cron jobs and tasks.
     • Endpoint Controller: Links services and pods.
     • Service Account Controller: Manages default service accounts.

5. Cloud Controller Manager (CCM)

   • Interfaces with cloud services, not available in Minikube.

Node Components

1. Kubelet

   • Agent running on each worker node.
   • Ensures pod specifications are met.
   • Manages container execution.

2. Kube-Proxy

   • Network component facilitating communication within pods and namespaces.
   • Interacts with API Server for network requests.

Container Runtime

• ContainerD:
  • Software running inside each container.
  • Manages operations in the latest Kubernetes version (1.27).

Summary

• Two main component types: Control Plane (Brain) and Worker Nodes.
• Control Plane has five components; Worker Nodes have two.
• All components function on Container Runtime.

Next Steps

• Future sessions will explore pods, deployments, and stateful sets.
• Complete documentation available in the linked repository.

Conclusion

• Lectures will continue to cover practical aspects of Kubernetes.
• Engagement encouraged through comments and feedback.