CKA Day 5: What is Kubernetes — Kubernetes Architecture Explained

Day 5/40 - What is Kubernetes - Kubernetes Architecture Explained

Overview

This video is Day 5 of the 40-day CKA preparation course. It delivers the foundational architecture deep-dive that every CKA candidate must master. The instructor breaks down the Control Plane (the brain) and Worker Nodes (the muscle), explaining how each component communicates, what it does, and why it matters for the exam. This is the most critical conceptual lesson for the ~25% “Cluster Architecture, Installation & Configuration” CKA domain.

Source Details

• Channel: Tech Tutorials with Piyush
• Playlist: 40 Days of Kubernetes (CKA Full Course)
• Companion Repository: piyushsachdeva/CKA-2024
• Challenge: #40daysofKubernetes
• Kubernetes Version: 1.30.2

Key Takeaways

1. The Two Halves of Kubernetes

Kubernetes clusters are divided into two logical halves:

Half	Role	Components
Control Plane	The brain — makes global decisions, detects & responds to cluster events	API Server, etcd, Scheduler, Controller Manager, Cloud Controller Manager
Worker Nodes	The muscle — runs the actual application workloads	kubelet, kube-proxy, Container Runtime

2. Control Plane Components (The Brain)

kube-apiserver (API Server)

• The front door of Kubernetes. All commands (kubectl), internal components, and external tools talk to the cluster through the API Server.
• Validates every request (authentication, authorization, admission controllers).
• The only component that talks directly to etcd.
• CKA Tip: If the API Server is down, the cluster is effectively frozen — no new pods, no updates, no scaling.

etcd

• The distributed key-value store that holds the entire cluster state: pod specs, node statuses, config maps, secrets, RBAC policies.
• Uses the Raft consensus algorithm for consistency and fault tolerance.
• CKA Tip: Backup etcd before upgrades. etcdctl snapshot save is a must-know command.

kube-scheduler

• Watches for newly created pods with no assigned node and selects the best node to run them.
• Considers: resource requests/limits, taints/tolerations, affinity/anti-affinity, data locality.
• CKA Tip: You can customize scheduling with PriorityClass, PodTopologySpread, and custom schedulers.

kube-controller-manager

n- Runs controller loops that watch the cluster state and drive actual state toward desired state.

• Key controllers: Node Lifecycle, Replication, Endpoints, Service Account & Token.
• CKA Tip: If a pod disappears, the Replication Controller creates a new one. If a node goes down, the Node Controller marks it NotReady and reschedules pods.

cloud-controller-manager

• Bridges Kubernetes with the underlying cloud provider (AWS, GCP, Azure).
• Manages cloud-specific resources: load balancers, storage volumes, node routes.
• CKA Tip: On-prem clusters may not have this; cloud-managed clusters (EKS, GKE, AKS) always do.

3. Worker Node Components (The Muscle)

kubelet

• The node agent that registers the node with the cluster and ensures containers are running in a Pod.
• Receives PodSpecs from the API Server and instructs the Container Runtime to create/maintain/destroy containers.
• CKA Tip: If kubelet is down on a node, that node appears NotReady; existing pods keep running but no new ones are scheduled.

kube-proxy

• The network proxy that maintains network rules on each node for Pod-to-Service communication.
• Implements Services using iptables or IPVS rules.
• CKA Tip: If kube-proxy fails, Services stop routing traffic to backend pods.

Container Runtime

• The software that actually runs containers. Kubernetes supports any runtime that implements the Container Runtime Interface (CRI).
• Common runtimes: containerd (default in modern clusters), CRI-O (Red Hat), Docker (deprecated as direct runtime, use containerd instead).
• CKA Tip: crictl is the CLI to inspect containers when debugging runtime issues.

4. How the Components Talk

User ──► kubectl ──► kube-apiserver ──► etcd
                           │
                           ├──► kube-scheduler (assigns node)
                           │
                           ├──► kube-controller-manager (maintains desired state)
                           │
                           └──► kubelet (on worker node) ──► Container Runtime
                                    │
                                    └──► kube-proxy (network rules)

5. Pods vs. Containers

• A Pod is the smallest deployable unit in Kubernetes.
• A Pod can contain one or more containers that share networking and storage.
• Containers within a Pod communicate via localhost and share a single IP address.
• CKA Tip: You rarely create individual containers in K8s — you create Pods (usually via Deployments).

6. What Is a Kubernetes Cluster?

• A set of Control Plane nodes (usually 1 or 3 for HA) plus Worker Nodes (any number).
• The Control Plane can run on dedicated machines or be co-located on worker nodes (not recommended for production).
• CKA Tip: For the exam, you’ll work with a single control plane + 2 worker nodes.

Cross-References

• Kubernetes Architecture
• Why Kubernetes?
• CKA Certification Overview
• CKA Progress Tracker

---

Ingested: 2026-05-21