Skip to content

Deploy an EKS cluster and get its Terraform code in minutes

TLDR; AWS EKS + Terraform + CloudSkiff do the job!
In this article we'll explain how to spin up an AWS EKS cluster in 1 min of work, and get Terraform code out of it for reproducibility and easy cleanup, with CloudSkiff, a CI/CD for infrastructure as code.

Deploy an EKS cluster

Why deploying an EKS cluster and gettin its Terraform code in minutes is tedious.​

If you need to deploy an EKS cluster and get its Terraform code in minutes this article explains how to do it using CloudSkiff.

Setting up new environments in EKS is a little tedious, and requires a lot of point and click work if you do it through the console.

Plus if something messes up, or you just want to shut it all down, you end up with a shitload of work cleaning up your AWS account and getting rid of now useless services. AWS didn’t make that simple (who designed that CLI again? And no, you can’t delete your VPC, there’s a NAT gateway attached to it. Did I mention that there is no automated cleanup function either?).

Enters Terraform

Describe everything as Terraform code, and you get a really easy way to deploy your new dev environment, a way that is reproducible and easy to clean up. And it makes it simpler to do things cleanly, with your environment neatly set up in a VPC for isolation.

Writing, optimizing and running Terraform code is a little tricky, and if you have your infra described as code, you might as well manage it in a CI/CD system like any other code. Right?

That’s why CloudSkiff is a platform for Infrastructure as Code that:

  • Day 1: makes getting started with infrastructure as code more approachable.
  • Day 2+: streamlines versioning, acts as the central place for automation, and enables collaboration around your templates and deployments .

We’re talking about AWS here, but CloudSkiff connects to other cloud providers too.

Workflow Infrastructure as code

Remember that you need to deploy an EKS cluster and get its Terraform code..​. So let’s dive into it. Start the timer, and let’s see how we launch a small dev cluster in 2 min of work. CloudSkiff will also generate basic but clean Terraform code for you, that you can then reuse and upgrade to evolve your environment.

1 - Create a CloudSkiff account

Easy, it’s here.

2 - Create a cloudSkiff IAM user in your AWS account

Sign into the AWS management console, then create a new aws IAM user for CloudSkiff. I called mine ClousdSkiff

Deploy an EKS cluster

Hit Add user then select programmatic access

Deploy an EKS cluster

We will create a new set of policies for this user to secure things up.

Deploy an EKS cluster

CloudSkiff needs access to EKS, EC2 and IAM. I created an easy, copy paste friendly permission set right there.

    "Version": "2012-10-17",
    "Statement": [
            "Effect": "Allow",
            "Action": [
            "Resource": "*"
            "Effect": "Allow",
            "Action": [
            "Resource": [

You don’t need to add tags.

Let’s create this user now. Once you’ve created your user, save your access key and secret key, we’ll need them soon.

Deploy an EKS cluster

3 - Add AWS to CloudSkiff

Great! We’ve created a new IAM cloudskiff user. Now let’s grant the CloudSkiff platform access using that user.

  1. Open the CloudSkiff app
  2. Navigate to the Integrations tab on the left
  3. Select AWS
  4. Enter your credentials, select your favorite AWS region
  5. Save. Keep the keys handy, we’ll need them later to configure our local AWS profile.

Deploy an EKS cluster

4 - Add permissions on a new infra as code Github repo

CloudSkiff will generate Terraform code for your infrastructure and save it in your repo. So we need to create a github repo that we want it to push to.

  1. Create a new private Github repository. Let’s call it cloudskiff-dev-eks
  2. Go to CloudSkiff’s integrations tab and select Github
  3. Connect your Github account

Note: CloudSkiff only needs access to the specific repo where you Terraform code will be stored.

5 - Cool. Let's deploy an EKS cluster!

The setup is complete. You’ll only have to do steps 1,2,3,4 once.

Now let’s see how we can launch an EKS cluster. Move to the CloudSkiff dashboard. That’s where you will monitor all your clusters, and launch new ones. Hit New Project.

Deploy an EKS cluster
  1. Select Templates. Templates are preconfigured EKS cluster that help you get started. You still have access to the Terraform behind it.
  2. Pick a name for your project
  3. Select AWS as the provider
  4. Select your usual region
  5. We’ll deploy a small cluster of t3.nano scaling between 1 and 3 machines. You can always come back to it and launch something more serious afterwards 🙂
  6. Enter your ssh public key (cat ~/.ssh/ to get it real quick on most systems). It should look like ssh-rsa BLABLABLA .
  7. Select your brand new cloudskiff-dev-eks repo
  8. Hit Save

You should land back to your dashboard, and tadaaaam: our project is there.

9. Hit Deploy! Our project will start, and we can monitor the progress in the Logs.

Deploy an EKS cluster

6 - Relax and check out the Terraform code we've generated

See that github logo? Hit it and you’ll land on your cloudskiff-dev-eks repository. The Terraform code that is executing right now has been stored on that repo. That means it is versioned, traced, and in case there is trouble you can roll back to older versions. GitOps become easier.

Deploy an EKS cluster

Meanwhile, AWS is doing its thing, starting the EKS cluster, VPCs, autoscaling groups described in this Terraform.

I am guessing you already are using AWS routinely, and you have the AWS CLI setup. Let’s take a look at that.

7 - Setup your local environment

All you need to do is :

  1. Create a CloudSkiff profile in your AWS credentials file, so that you can access your machines with your cloudskiff IAM.

aws_access_key_id = ..#you probably already have something here
aws_secret_access_key = #here too

#Create this
aws_access_key_id = .. 
aws_secret_access_key = .. # told you we'd need that later 

2. Set your local environment variables $KUBECONFIG and $AWS_PROFILE . KUBECONFIG should contain the path to your kubeconfig file.

We will download it from CloudSkiff later, so let’s just prepare a folder to save it, for example $HOME/code/cloudskiff/config/aws and define the KUBECONFIG to point to the file.

export AWS_PROFILE=cloudskiff;

#the path to where you will store your kubeconfig file export


8 - Connect

Wait a few minutes (10–15) for AWS to assign resources. At some point, your project will be deployed (you will see a green ball in the UI).

1.Get the kubeconfig from the CloudSkiff dashboard.

2. Rename it to kubeconfig-dev-cluster.Then move it to ~/code/cloudskiff/config/kubeconfig-dev-cluster or the place of your choosing, as long as it matches your $AWS_PROFILE

3. Check: echo $AWS_PROFILE; echo $KUBECONFIG . It should output something like that:

AWS_PROFILE: cloudskiff 
KUBECONFIG: ~/code/cloudskiff/config/aws/kubeconfig-dev-cluster 

4. Now run kubectl get nodes , or k9s if you prefer. You’re in! Cluster deployed!

(9 - Destroy)

To destroy your cluster and cleanup everything, well: just hit the Destroy button on CloudSkiff. Everything will be cleaned up automatically and ready for a re-deploy!

Reading this, you might think I took more than 2 min, because I sprayed screenshots everywhere.

Thinking about it, most of the things I did were just one-off for setup:

  1. (only once) Create a CloudSkiff account
  2. (only once) Create an IAM
  3. (only once) add it to CloudSkiff
  4. (only once) add permissions on a new infra as code Github repo
  5. Select 5 options
  6. Press a Deploy button
  7. (only once) make sure my local AWS profile was configured
  8. Get a kubeconfig
  9. Connect!

Only 5, 6, 8, 9 are steps you need to do for each deployment, and they are mostly buttons to press or single lines of command.

I hope you liked that! We haven’t looked in detail in the terraform code together, so I will keep that for an upcoming post.

Key metrics for infrastructure automation

Check your progress in your infrastructure automation journey with metrics that matter.

Maintain reliability and security while moving at developers speed.

Track inconsistencies that were invisible so far.