Module 5: Deployments & Services

🔹 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:

1. ClusterIP (default) – Exposes the service inside the cluster only.

2. NodePort – Exposes the service on a static port on each Node.

3. 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

Explanation:

• Targets Pods with label app=nginx.

• Exposes port 80 inside the cluster.

• NodePort assigns a static port (e.g., 30007) on each Node.

---

Accessing the App

• Inside cluster: curl http://nginx-service

• Outside cluster (NodePort): http://<NodeIP>:<NodePort>

• On Cloud (LoadBalancer): Public LoadBalancer IP is provided.

---

🔹 Summary

• Deployments = Best way to run production apps (scaling, rolling updates, rollback).

• Services = Stable network endpoint for Pods.

• Together, they form the backbone of Kubernetes workloads.

---