Kubernetes Deployments & Services Explained

🔹 Introduction

We explored Pods and ReplicaSets. While these are foundational concepts, real-world Kubernetes applications rarely use Pods or ReplicaSets directly. Instead, we rely on Deployments to manage them, and Services to make our applications accessible.

This module will cover:

What is a Deployment?
Benefits of using Deployments over Pods/ReplicaSets
Rolling updates & rollback with Deployments
What is a Service in Kubernetes?
Different types of Services (ClusterIP, NodePort, LoadBalancer)
Hands-on example: Deploying an Nginx app and exposing it with a Service

🔹 Deployments in Kubernetes

A Deployment is a higher-level abstraction in Kubernetes that:

Manages ReplicaSets automatically.
Ensures the desired state of Pods (self-healing).
Provides rolling updates and rollbacks.
Makes scaling applications easier.

Why not just use Pods or ReplicaSets?

Pods are ephemeral (they die if a node fails).
ReplicaSets provide scaling but lack deployment strategies.
Deployments combine both — scaling, self-healing, rolling updates, and rollback.

Example: Nginx Deployment

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:latest
        ports:
        - containerPort: 80

Explanation:

Runs 3 replicas of the Nginx Pod.
Ensures if one Pod fails, another is created.
Can be updated easily with a new image version.

Scaling Deployments

kubectl scale deployment nginx-deployment --replicas=5

Now, Kubernetes will automatically adjust the number of Pods to 5.

Rolling Updates

If you update the Deployment to use nginx:1.25, Kubernetes will:

Gradually replace old Pods with new ones.
Ensure zero downtime.

Rollback if needed:

kubectl rollout undo deployment nginx-deployment

🔹 Services in Kubernetes

Deployments create Pods, but Pods get dynamic IPs that change if a Pod restarts. 👉 That’s where Services come in!

A Service is a stable network endpoint that routes traffic to a set of Pods.

Types of Services:

ClusterIP (default) – Exposes the service inside the cluster only.
NodePort – Exposes the service on a static port on each Node.
LoadBalancer – Creates a cloud load balancer (e.g., AWS ELB, GCP LB).

Example: Exposing Nginx with a Service

apiVersion: v1
kind: Service
metadata:
  name: nginx-service
spec:
  selector:
    app: nginx
  ports:
    - protocol: TCP
      port: 80
      targetPort: 80
  type: NodePort