Guides

eksctl create cluster: Create an Amazon EKS Cluster Step by Step (and Delete It)

By Ravindra Bagale, AWS & DevOps trainer · LinkedIn · Published: 16 October 2026 · Last updated: 17 October 2026

eksctl create cluster creates a working Amazon EKS cluster in one command: a VPC, the EKS control plane, a managed node group of EC2 workers, and your kubectl config. The best way is a small YAML file and eksctl create cluster -f cluster.yaml, which takes about 15–20 minutes. Remember that the control plane costs 0.10 USD per hour plus the EC2 nodes, so delete the cluster with eksctl delete cluster when your lab is done.

eksctl create cluster --name lab --region ap-south-1 --version 1.30 \
  --nodegroup-name ng-1 --node-type t3.medium --nodes 2 --managed
kubectl get nodes
eksctl delete cluster --name lab --region ap-south-1

A live example first

We create a small lab cluster in Mumbai with a YAML file, so it is easy to repeat and to delete.

1. Check the tools:

aws sts get-caller-identity
eksctl version
kubectl version --client

2. Write cluster.yaml:

apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
  name: rb-lab
  region: ap-south-1
  version: "1.30"
managedNodeGroups:
  - name: ng-general
    instanceType: t3.medium
    desiredCapacity: 2
    minSize: 1
    maxSize: 3
    volumeSize: 20
    labels: { role: general }
    tags: { project: rb-lab }

3. Create it:

eksctl create cluster -f cluster.yaml

[ℹ]  eksctl version 0.19x.0
[ℹ]  using region ap-south-1
[ℹ]  building cluster stack "eksctl-rb-lab-cluster"
[ℹ]  deploying stack "eksctl-rb-lab-cluster"
...
[ℹ]  building managed nodegroup stack "eksctl-rb-lab-nodegroup-ng-general"
[✔]  saved kubeconfig as "/home/ravi/.kube/config"
[✔]  EKS cluster "rb-lab" in "ap-south-1" region is ready

4. Test it:

kubectl get nodes
kubectl create deployment web --image=nginx:1.27
kubectl expose deployment web --type=LoadBalancer --port=80
kubectl get svc web -w

After a minute or two, EXTERNAL-IP shows a load balancer DNS name, and the nginx page opens. Friends, that is a real, production-style Kubernetes cluster, made from 15 lines of YAML.

What does eksctl create?

You can see these stacks in the CloudFormation console. Do not edit them by hand. Change the YAML and use eksctl.

How do I create a cluster step by step?

1. Install the AWS CLI and run aws configure (or use an SSO profile).
2. Install kubectl and eksctl.
3. Write cluster.yaml with name, region, version and a node group.
4. Run eksctl create cluster -f cluster.yaml --dry-run to see the full config.
5. Run eksctl create cluster -f cluster.yaml.
6. Run kubectl get nodes and check that they are Ready.
7. Install add-ons you need, for example the EBS CSI driver for persistent volumes.
8. Delete the cluster when done.

Install eksctl on each system

eksctl was first created by Weaveworks and is now maintained as an official tool for Amazon EKS. Follow the install page in the EKS user guide for the current method.

Useful variations

# Use an existing VPC
vpc:
  subnets:
    private:
      ap-south-1a: { id: subnet-0aaa... }
      ap-south-1b: { id: subnet-0bbb... }

# Cheaper lab nodes with Spot
managedNodeGroups:
  - name: ng-spot
    instanceTypes: ["t3.medium", "t3a.medium"]
    spot: true
    desiredCapacity: 2

# Pod IAM roles (IRSA)
iam:
  withOIDC: true

Other handy commands:

eksctl get cluster --region ap-south-1
eksctl scale nodegroup --cluster rb-lab --name ng-general --nodes 3
eksctl create addon --cluster rb-lab --name aws-ebs-csi-driver
eksctl utils write-kubeconfig --cluster rb-lab

Give other people access

The IAM identity that created the cluster is the only admin at first. To add a teammate's role:

aws eks create-access-entry --cluster-name rb-lab --principal-arn arn:aws:iam::123456789012:role/devops-team
aws eks associate-access-policy --cluster-name rb-lab \
  --principal-arn arn:aws:iam::123456789012:role/devops-team \
  --policy-arn arn:aws:eks::aws:cluster-access-policy/AmazonEKSClusterAdminPolicy \
  --access-scope type=cluster

Give admin only to people who need it. Use the view or edit policies for others.

Cost and clean-up

To delete cleanly:

1. kubectl delete svc --all (removes load balancers).
2. kubectl delete pvc --all (removes EBS volumes).
3. eksctl delete cluster -f cluster.yaml --wait.
4. Check CloudFormation and EC2 → Load balancers and Volumes for leftovers.

Other clouds

• Google GKE: gcloud container clusters create-auto lab --region asia-south1.
• Azure AKS: az aks create -g rb-rg -n lab --node-count 2.

Real example

At KubeCon Europe 2019, Spotify engineers gave a talk about how they accidentally deleted Kubernetes clusters during their migration to GKE, and how they recovered with no visible effect for users. One lesson they shared was to manage clusters as code, so a lost cluster can be recreated. eksctl YAML files kept in Git give you the same safety: the cluster can be rebuilt with one command.

Mistakes to avoid

1. Leaving the lab cluster running. It costs money every hour.
2. Deleting the cluster before LoadBalancer Services. The VPC deletion can get stuck on leftover load balancers.
3. Editing CloudFormation stacks by hand. eksctl gets out of sync.

Summary table