Skip to content
Latchkey

What Is Amazon EKS? Managed Kubernetes on AWS

Amazon EKS (Elastic Kubernetes Service) is AWS managed Kubernetes: AWS runs the control plane and you run workloads on worker nodes or Fargate.

Kubernetes is powerful but operationally heavy to run yourself. EKS removes the hardest part by managing the Kubernetes control plane for you, with high availability and automatic patching. You bring worker capacity and your manifests; AWS keeps the cluster brain healthy.

What EKS manages for you

EKS runs the Kubernetes API server, scheduler, and etcd datastore across multiple availability zones. You no longer babysit the control plane. You do still choose how worker nodes run: managed node groups, self-managed nodes, or serverless Fargate profiles.

How you interact with it

You talk to EKS the same way you talk to any Kubernetes cluster: with kubectl and standard manifests or Helm charts. EKS authenticates kubectl using IAM, mapping AWS identities to Kubernetes roles, so access control ties back into your AWS account.

When EKS makes sense

  • You already run Kubernetes and want it managed on AWS.
  • You need the broad Kubernetes ecosystem of operators and tooling.
  • You run many services that benefit from shared cluster scheduling.
  • You want portability across clouds using a common Kubernetes API.

EKS vs ECS

ECS is simpler and AWS-specific; EKS gives you the full Kubernetes ecosystem at the cost of more complexity. Teams without existing Kubernetes investment often start on ECS or Fargate; teams standardized on Kubernetes pick EKS.

Role in CI/CD

A pipeline builds an image, pushes it to ECR, then applies updated manifests or a Helm release to the cluster. Authentication uses an IAM/OIDC role that maps to a Kubernetes service account. GitOps tools like Argo CD can also watch a repo and reconcile the cluster, letting CI just update image tags in Git.

Key takeaways

  • EKS is managed Kubernetes: AWS runs the control plane, you run workloads.
  • You use standard kubectl and Helm, with IAM mapped to Kubernetes access.
  • Pipelines deploy by pushing to ECR and applying manifests or a Helm release.

Related guides

Tired of flaky CI? Latchkey auto-heals failed jobs and retries them for you. Start free →