Understanding Amazon EKS: Simplifying Kubernetes on AWS

---

Introduction

If you are learning Kubernetes today, you’ll hear a lot about EKS.

Most tutorials say:

• “EKS is managed Kubernetes”
• “EKS runs on AWS”

But they rarely explain what that actually means in practice.

This blog explains EKS in simple terms:

• What it does
• What AWS manages vs what you manage
• How it fits into real-world DevOps

---

1. What is Amazon EKS?

Amazon EKS (Elastic Kubernetes Service) is:

A managed service that runs Kubernetes for you on AWS.

Instead of setting up Kubernetes yourself, AWS provides:

• Control Plane
• High availability
• Integration with AWS services

You just focus on running workloads.

---

2. Why EKS exists

Running Kubernetes manually is not easy.

You need to manage:

• API Server
• etcd
• Scheduler
• Controller Manager

With EKS:

👉 AWS manages all of this for you

This reduces:

• Setup complexity
• Maintenance effort
• Risk of control plane failures

---

3. EKS architecture (simple view)

EKS still follows Kubernetes architecture.

But responsibilities are split.

---

✅ AWS manages (Control Plane)

AWS handles:

• API Server
• etcd
• Scheduler
• High availability

👉 You don’t see or access these directly

---

✅ You manage (Data Plane)

You handle:

• Worker nodes (EC2 or Fargate)
• Pods and deployments
• Networking and scaling decisions

---

Simple mental model

AWS → Control Plane
You → Nodes + Workloads

---

4. Two ways to run workloads in EKS

---

Option 1: EC2 Nodes

You create and manage nodes (VMs):

• Full control
• More flexible
• Requires maintenance

Example command:

kubectl get nodes

---

Option 2: Fargate (serverless)

AWS runs Pods for you without managing nodes.

• No server management
• Easier setup
• Less flexibility

---

5. How EKS workflow looks

Typical flow:

1. Create EKS cluster
2. Add worker nodes (EC2/Fargate)
3. Deploy applications using kubectl
4. Kubernetes schedules Pods on nodes

Example:

kubectl apply -f deployment.yaml

👉 Same as normal Kubernetes — nothing changes at usage level

---

6. What EKS simplifies

EKS removes the hardest part of Kubernetes:

---

✅ Control Plane management

• No need to install Kubernetes components
• No need to maintain etcd
• No need to handle upgrades manually

---

✅ High availability

AWS runs control plane across:

• Multiple AZs

👉 Reduces risk of cluster failure

---

✅ AWS integration

EKS integrates with:

• IAM (security)
• ELB (load balancing)
• EBS (storage)

👉 Makes production usage easier

---

7. What EKS does NOT manage

This is important.

Even with EKS, you still manage:

• Node scaling (EC2 mode)
• Application deployments
• Pod configurations
• Debugging issues

👉 EKS does NOT eliminate DevOps work
👉 It just simplifies infrastructure layer

---

8. Common beginner misunderstanding

---

❌ “EKS runs everything”

Reality:

• AWS manages control plane
• YOU manage workloads

---

❌ “EKS removes Kubernetes complexity”

Reality:

• Kubernetes is still Kubernetes
• Debugging Pods is unchanged

---

❌ “Fargate is always better”

Reality:

• Fargate is easier
• But less flexible

---

9. Real-world use cases

EKS is commonly used for:

• Microservices applications
• Scalable APIs
• Cloud-native workloads
• CI/CD pipelines

---

10. When should you use EKS?

Use EKS if:

• You want managed Kubernetes
• You are already using AWS
• You don’t want to manage control plane

---

11. When EKS may not be needed

Avoid EKS if:

• You are learning Kubernetes locally
• You need full control over cluster internals
• You want cheaper lightweight setups

---

12. Key commands remain the same

Even in EKS, Kubernetes commands remain identical:

kubectl get pods
kubectl get nodes
kubectl get svc
kubectl describe pod <pod-name>

👉 This is important:

EKS is still Kubernetes — just managed differently

---

🚀 InfraDecode Takeaway

EKS is Kubernetes with AWS managing the control plane for you.
You still own the nodes, workloads, and debugging responsibilities.
Think of EKS as reducing infrastructure complexity, not Kubernetes complexity.
Understanding core Kubernetes is still essential even when using EKS.