Deploying a Containerized App on AWS using Amazon EKS

Topics: Containers, Kubernetes, EKS, AWS

Overview

This mini-project demonstrates container orchestration on AWS using Amazon Elastic Kubernetes Service (EKS). You'll create a managed Kubernetes cluster, add worker nodes, deploy a containerized Nginx application, and expose it to the internet using a LoadBalancer service. This hands-on project covers the complete workflow of container deployment in the AWS cloud environment.

Key Concepts

Concept	Description
Amazon EKS	Managed Kubernetes service that runs containerized applications on AWS
Managed Node Groups	AWS-managed EC2 instances that serve as worker nodes in the cluster
kubectl	Command-line tool for interacting with Kubernetes clusters
eksctl	AWS CLI tool for creating and managing EKS clusters
Kubernetes Deployment	Declarative way to manage containerized applications
LoadBalancer Service	Kubernetes service type that exposes applications via AWS ELB
CloudShell	Browser-based shell environment for AWS management

Prerequisites

• Active AWS account with billing enabled
• Root account access or IAM permissions for EKS, EC2, VPC, and IAM services
• Basic understanding of containers and Kubernetes concepts
• Region: Mumbai (ap-south-1) or consistent with your account setup

Architecture Overview

Click to expand Architecture Diagram

Create EKS Cluster and Deploy Application

1. Access CloudShell and Verify Tools:

   • AWS Console (Mumbai region) → Click CloudShell icon
   • Terminal opens in browser
   • Verify installed tools:

     aws --version
     kubectl version --client
     eksctl version

   NOTE

   If eksctl is missing, CloudShell usually has it pre-installed. You can install it if needed using: curl --silent --location "https://github.com/weaveworks/eksctl/releases/latest/download/eksctl_$(uname -s)_amd64.tar.gz" | tar xz -C /tmp && sudo mv /tmp/eksctl /usr/local/bin

2. Create EKS Cluster and Node Group:

   eksctl create cluster \
   --name eks-lab-cluster \
   --region ap-south-1 \
   --nodes 2 \
   --node-type t3.medium \
   --managed

   Wait for completion.

   TIP

   Troubleshooting IAM permissions: If the command fails, verify your credentials:

   aws sts get-caller-identity

   The output should show your root account ID or an ARN with proper permissions. Ensure you have EKS, EC2, VPC, and IAM permissions.

   NOTE

   What this command creates:

   • New VPC with proper subnets and routes
   • EKS cluster control plane (managed by AWS)
   • Managed Node Group with 2 t3.medium EC2 worker nodes
   • Automatic kubeconfig configuration for kubectl access

3. Verify Cluster and Nodes:

   kubectl get nodes

   Expected: 2 nodes in Ready state.

4. Deploy Nginx Application:

   kubectl create deployment webapp --image=nginx
   kubectl get pods

5. Expose Application with LoadBalancer:

   kubectl expose deployment webapp --type=LoadBalancer --port=80
   kubectl get svc

   Wait until EXTERNAL-IP becomes a value.

6. Test and Access Application:

   • Copy the EXTERNAL-IP from the service output
   • Open in browser: http://<EXTERNAL-IP>
   • Expected: Nginx welcome page

Validation

Validation

• CloudShell Access: Terminal successfully opened in browser
• Tools Verification: aws, kubectl, and eksctl commands work
• EKS Cluster: Cluster creation completed successfully
• Node Group: 2 worker nodes created and in Ready state
• Nginx Deployment: Pod created and running
• LoadBalancer Service: Service created with external IP
• Application Access: Nginx welcome page loads in browser

Cost Considerations

Cost Considerations

• EKS Control Plane: $0.10 per hour (always running)
• EC2 Worker Nodes: t3.medium pricing (~$0.04/hour each)
• Elastic Load Balancer: $0.0225 per hour + $0.008 per GB data
• Data Transfer: Standard AWS data transfer rates
• Free Tier: Limited EKS free tier available
• Estimated Cost: $5-15 for 1-2 hour lab session

Cleanup

Cleanup

1. Delete LoadBalancer service:

   kubectl delete svc webapp

2. Delete deployment:

   kubectl delete deployment webapp

3. Delete EKS cluster (this removes everything):

   eksctl delete cluster --name eks-lab-cluster --region ap-south-1

4. Verify all resources are removed to avoid charges

Result

Successfully deployed a containerized Nginx application on Amazon EKS using managed Kubernetes services. Demonstrated the complete container orchestration workflow from cluster creation through application exposure. Mastered EKS cluster management, kubectl operations, and AWS LoadBalancer integration for containerized applications.

Viva Questions

1. What is the difference between Amazon EKS and self-managed Kubernetes?
2. Why use managed node groups instead of self-managed nodes?
3. How does a LoadBalancer service work in Kubernetes?
4. What are the benefits of using CloudShell for AWS management?