What Is Amazon EKS (AWS EKS)? A Guide to Managed Kubernetes

Sunucun 3 min read

Amazon EKS (Elastic Kubernetes Service) is the managed Kubernetes service of Amazon Web Services (AWS). Kubernetes is an open-source platform that automates the deployment, scaling and management of containerized applications. In EKS, AWS runs the hardest part of Kubernetes to operate, the control plane, so you can focus on the nodes your applications run on and on your workloads.

Illustration: Kubernetes

What Does AWS Manage in EKS, and What Do You Manage?

According to the AWS documentation, EKS is used in two main ways:

  • EKS standard: When you create a cluster, AWS manages the Kubernetes control plane (the API server, scheduler and the components that store cluster state). The nodes your applications run on are managed by you.
  • EKS Auto Mode: AWS extends management to the nodes as well; choosing suitable EC2 instances, scaling, OS patching and cost optimization happen automatically.

EKS is certified Kubernetes-conformant, so you can use your standard Kubernetes applications and community tools such as kubectl and Helm without changes.

Node (Compute) Options

OptionHow it worksWhen it fits
EKS Auto ModeAWS creates, scales and patches nodes automaticallyTeams that don’t want to manage nodes
Managed node groupsYou manage EC2 instances as groups with AWS’s helpThose who want control over instance types and settings
Self-managed nodesYou configure EC2 instances entirely yourselfThose who need a custom OS or configuration
AWS FargateNo servers to manage; pay for vCPU and memory per podVariable or small workloads
EKS Hybrid Nodes / EKS AnywhereUse servers in your own data centerThose connecting on-premises infrastructure to Kubernetes

EKS Pricing and Kubernetes Version Support

  • Cluster fee: USD 0.10 per cluster per hour for a Kubernetes version in standard support; USD 0.60 per hour in extended support.
  • Support periods: Each Kubernetes version gets 14 months of standard support after its release in EKS, followed by 12 months of extended support. Upgrading clusters regularly keeps costs down.
  • Billed separately: EC2 instances (or Fargate usage) for nodes, EBS volumes, public IPv4 addresses, load balancers and data transfer. With Auto Mode, an Auto Mode fee applies on top of the EC2 price.

Prices are based on AWS’s EKS pricing page and can change; check the current amounts before you deploy.

How to Create Your First EKS Cluster

You can create a cluster with the AWS Console, AWS CLI, eksctl, CloudFormation or Terraform. The quickest way is eksctl, the command-line tool commonly used with EKS:

# Example: a cluster with Auto Mode enabled
eksctl create cluster --name demo-cluster --region eu-central-1 --enable-auto-mode

# Connect to the cluster and list nodes and pods
kubectl get nodes
kubectl get pods -A

Once the cluster is ready, deploy your applications with kubectl apply or Helm. Access is managed with both Kubernetes and AWS IAM; container images are usually stored in Amazon ECR, and CloudWatch or Prometheus is used for monitoring.

When Does EKS Make Sense?

  • Makes sense: When you have many microservices, your team knows Kubernetes and you want deep integration with AWS services (RDS, S3, IAM, load balancers).
  • Can be overkill: For a single website or a few containers, the learning and operating overhead of Kubernetes is often unnecessary. In that case Docker or Docker Compose on a single server is much simpler and cheaper.

If you want to run your containers on a fixed-cost server with root access, take a look at our VDS server options. Other articles in our AWS series: Amazon EC2, Amazon RDS, AWS IAM and Amazon VPC.

Conclusion

Amazon EKS hands the Kubernetes control plane over to AWS so you can operate clusters more easily, and Auto Mode automates node management too. Considering the cluster fee, version support periods and node costs together, EKS is a strong option for growing applications that need Kubernetes, and often more complex than necessary for small projects.

Categories

Latest guides