ReplicaSet & Deployment

January 29, 2026

📌 Kubernetes ReplicaSet & Deployment

🔹 What is a ReplicaSet?

A ReplicaSet (RS) is a Kubernetes controller that ensures a specified number of identical Pods are running at all times.

If a Pod:

Crashes
Is deleted
Node goes down

➡️ ReplicaSet automatically creates a new Pod to maintain the desired count.

🔹 Why ReplicaSet is Used

ReplicaSet is used to:

Maintain high availability
Ensure fault tolerance
Automatically heal Pods
Run multiple replicas of the same application

🔹 How ReplicaSet Works

User defines desired number of replicas
ReplicaSet uses labels & selectors
It continuously compares:
- Desired state vs Actual state
If mismatch occurs → it creates or deletes Pods

🔹 ReplicaSet Key Concepts

1️⃣ Replicas

Number of Pod copies to run
Defined in spec.replicas

2️⃣ Pod Template

Blueprint for Pods
All Pods are identical
Defined under spec.template

3️⃣ Labels & Selectors

RS identifies Pods using labels
Selector must match Pod labels (mandatory)

⚠️ Mismatch will cause RS to fail

🔹 ReplicaSet YAML Structure (Important)


spec:
  replicas: 3
  selector:
    matchLabels:
      app: paytm
  template:
    metadata:
      labels:
        app: paytm

🔹 What ReplicaSet Can Do

✅ Auto-heal Pods
✅ Scale up/down
✅ Maintain desired state
✅ Load balancing (via Service)

❌ ReplicaSet Limitations (Very Important)

❌ No rolling updates
❌ No rollback support
❌ No version tracking
❌ Updating image requires manual Pod deletion

⚠️ Not recommended for production directly

🔹 When to Use ReplicaSet

Rarely used directly
Mostly created and managed by Deployments
Useful for:
- Learning Kubernetes basics
- Understanding controllers

🔹 ReplicaSet Summary (Interview-Ready)

ReplicaSet ensures the desired number of Pods are always running using labels and selectors but lacks rolling update and rollback features.

📌 Kubernetes Deployment – Detailed Notes

🔹 What is a Deployment?

A Deployment is a higher-level Kubernetes controller that manages:

ReplicaSets
Pods
Application versions
Updates & rollbacks

It is the most commonly used workload in Kubernetes.

🔹 Why Deployment is Used

Deployment is used because it provides:

Zero-downtime deployments
Rolling updates
Easy rollback
Version control
Production-grade management

👉 Best choice for real-world applications

🔹 How Deployment Works (Flow)


Deployment → ReplicaSet → Pods

Deployment creates a ReplicaSet
ReplicaSet creates Pods
On update:
- New ReplicaSet is created
- Old ReplicaSet is scaled down gradually

🔹 Deployment Key Features

1️⃣ Rolling Updates

Updates Pods gradually
No downtime
Default strategy

2️⃣ Rollback

Revert to previous version
Uses revision history

3️⃣ Versioning

Every update creates a new revision
Stored in Deployment history

🔹 Deployment YAML Highlights


spec:
  replicas: 3
  selector:
    matchLabels:
      app: paytm
  template:
    spec:
      containers:
      - name: cont1
        image: myapp:1.0

Changing the image automatically triggers a rollout.

🔹 Deployment Strategies

🔄 RollingUpdate (Default)

Gradual replacement
Zero downtime


strategy:
  type: RollingUpdate

🔁 Recreate

Deletes all Pods first
Causes downtime


strategy:
  type: Recreate

🔹 Deployment Rollout Commands


kubectl rollout status deploy movies
kubectl rollout history deploy movies
kubectl rollout undo deploy movies
kubectl rollout undo deploy movies --to-revision=1

🔹 Scaling in Deployment

Manual scaling


kubectl scale deploy movies --replicas=10

Auto scaling (HPA)


kubectl autoscale deploy movies --cpu-percent=50 --min=2 --max=10

🔹 What Deployment Can Do

✅ Zero downtime upgrades
✅ Auto healing
✅ Rollback support
✅ Version history
✅ Auto scaling
✅ Production ready

🔹 When to Use Deployment

Use Deployment when:

Running stateless applications
You need rolling updates
You want rollback capability
Running apps in production

🔹 ReplicaSet vs Deployment (Key Comparison)

Feature	ReplicaSet	Deployment
Auto-heal	✅	✅
Scaling	✅	✅
Rolling update	❌	✅
Rollback	❌	✅
Versioning	❌	✅
Production use	❌	✅