Kubernetes Architecture

Kubernetes (K8s) Architecture & Deployments

Core Concepts

Kubernetes is a distributed system for managing container workloads. It consists of a Control Plane (brain) and Data Plane (worker nodes).

1. Control Plane vs Worker Nodes

• Control Plane: kube-apiserver (Entrypoint), etcd (Key-Value Store, the absolute source of truth), kube-scheduler (Assigns Pods to Nodes), kube-controller-manager.
• Worker Node: kubelet (Agent that ensures containers are running), kube-proxy (Network routing/iptables), Container Runtime (containerd).

2. Deployment Strategies

• Rolling Update: Replaces pods incrementally. Zero downtime.
• Blue/Green: Spins up a full duplicate environment (Green), switches traffic instantly at the Load Balancer level.
• Canary: Routes a small percentage (e.g., 5%) of traffic to the new version to monitor metrics before full rollout.

Architecture Map

%%{init: {"theme": "default", "flowchart": {"useMaxWidth": true}}}%%
flowchart TD
    subgraph ControlPlane ["Control Plane (Master)"]
        A["kube-apiserver"]
        B["etcd (State DB)"]
        C["kube-scheduler"]
        D["kube-controller-manager"]
    end
    
    subgraph WorkerNode ["Worker Node (Data Plane)"]
        E["kubelet (Node Agent)"]
        F["kube-proxy (Networking)"]
        G["Container Runtime (containerd)"]
        H["Pod 1"]
        I["Pod 2"]
    end
    
    A <--> B
    A <--> C
    A <--> D
    A <-->|"gRPC/TLS"| E
    E -->|"Instructs"| G
    G -->|"Runs"| H
    G -->|"Runs"| I
    F -.->|"Routes IP traffic"| H