Notes

1. Kubernetes Overview (What & Why)

2 min read

What is Kubernetes (K8s)?

Core Idea

Kubernetes (K8s) is a container orchestration system that automatically deploys, scales, and manages containerized applications (like [[Docker]] containers).

💡 Memory Trick: [[Docker]] = runs containers | Kubernetes = manages containers at scale.

🧩 The Problem it Solves

Before K8s, managing standalone containers at an enterprise level caused massive operational overhead:

  • ❌ Manual deployments.
  • ❌ No auto-scaling during traffic spikes.
  • ❌ No self-healing (if a container crashed, it stayed dead).
  • ❌ Complex networking and manual load balancing.

👉 Kubernetes solves all of this by automating the operational lifecycle.

🏗️ High-Level Architecture

A Kubernetes [[Cluster]] is divided into two primary components:

  1. Control Plane (The Brain 🧠):
    • Makes global cluster decisions.
    • Schedules workloads.
    • Constantly monitors and maintains the desired state.
  2. Worker Nodes (The Execution 💪):
    • The actual machines (VMs or physical servers) that run your applications.

⚙️ Key Capabilities

  • Container Orchestration: Runs and manages containers across multiple disparate machines as if they were one unified system.
  • Auto-Scaling: Automatically scales workloads up or down based on CPU, memory, or custom metrics.
  • Self-Healing: If a [[Pod]] dies, K8s automatically restarts or replaces it.
  • Load Balancing: Distributes incoming network traffic safely across available, healthy Pods.
  • Rolling Updates: Enables seamless, zero-downtime deployments when pushing new versions of an app.

🧠 Core Terminology

  • Cluster: The entire system/ecosystem working together.
  • Node: A single worker machine within the cluster.
  • Pod: The smallest, most basic deployable unit in Kubernetes.
  • Container: The actual application running inside the Pod.

💡 Insights & Real-World Application

Real-World DevOps Example

Imagine deploying a Machine Learning model (e.g., via SageMaker):

  • Without K8s: You rely on 1 server. If it crashes, your app goes down entirely.
  • With K8s: You declare that you want 3 pods running. If 1 crashes, K8s automatically replaces it and instantly shifts traffic to the healthy pods. This ensures production-grade reliability.

The Interview Answer

“Kubernetes is a container orchestration platform that automates the deployment, scaling, networking, and self-healing of containerized applications across a cluster of machines.”

CKA Concept: The Desired State

K8s fundamentally works on a declarative Desired State. You write a configuration file declaring, “I want 3 pods running.” K8s acts as an active control loop, continuously monitoring the cluster to ensure it always matches that declaration, no matter what underlying hardware fails.

Graph View

Backlinks