canonical: https://jentic.com/apis/amazonaws.com/amazon-eks

# AWS Amazon Elastic Kubernetes Service

Jentic publishes the only available OpenAPI specification for Amazon Elastic Kubernetes Service, keeping it validated and agent-ready. Amazon EKS is a managed Kubernetes service that runs upstream Kubernetes on AWS without requiring teams to operate the control plane themselves. The API covers cluster lifecycle management, managed node groups, Fargate profiles, EKS add-ons, and identity provider configuration. It is designed for platform teams that need to provision, scale, and upgrade Kubernetes clusters across multiple AWS regions while keeping pace with upstream Kubernetes releases.

## For AI agents

Provision and manage Amazon EKS Kubernetes clusters, node groups, Fargate profiles, and add-ons across AWS regions through a single API.

## Scope

Does not handle Kubernetes object operations, container image builds, or in-cluster RBAC policies - use for EKS control-plane and node-group management only.

## Capabilities

- Create and delete EKS clusters with custom VPC, logging, and Kubernetes version configuration
- Provision managed node groups with autoscaling settings, instance types, and AMI release versions
- Attach Fargate profiles to run pods on serverless compute without managing EC2 nodes
- Install, update, and remove EKS add-ons such as VPC CNI, CoreDNS, and kube-proxy
- Associate OIDC identity provider configurations to authenticate Kubernetes users against external IdPs
- Update cluster encryption configuration to enable envelope encryption of Kubernetes secrets
- List and describe cluster, node group, and add-on update history for change auditing

## Use cases

### Multi-Region Kubernetes Platform Provisioning

Platform engineering teams use the Amazon EKS API to stand up production-grade Kubernetes clusters across multiple AWS regions with consistent network, logging, and encryption configuration. The API drives cluster creation, node group sizing, and Fargate profile attachment so teams can codify cluster topology in pipelines instead of clicking through the console. A typical regional rollout creates one cluster, two node groups, and three Fargate profiles in under thirty minutes when run as an automated workflow.

Example prompt: Create an EKS cluster named prod-eu in eu-west-1 with Kubernetes version 1.28, then attach a managed node group of three t3.large nodes.

### Fargate Workload Onboarding

Teams migrating workloads to serverless containers use the EKS API to attach Fargate profiles that select pods by namespace and label. This removes the need to manage EC2 worker nodes for specific workloads while keeping the rest of the cluster on managed node groups. Fargate profile creation typically completes in two to three minutes and supports up to five selector rules per profile.

Example prompt: Create a Fargate profile on cluster prod-eu that selects pods in the batch namespace with label workload=etl.

### Add-on Lifecycle Management

EKS add-ons let teams keep cluster components such as VPC CNI, CoreDNS, and kube-proxy on supported, vendor-curated versions. The API handles install, version updates, and removal so cluster operators can roll out coordinated upgrades across many clusters. Add-on updates are non-disruptive when configured with the PRESERVE conflict resolution strategy.

Example prompt: Update the vpc-cni add-on on cluster prod-eu to version v1.15.1-eksbuild.1 with the PRESERVE conflict resolution strategy.

### Agent-Driven Cluster Operations via Jentic

AI agents use the EKS API through Jentic to handle routine cluster operations such as listing clusters, checking upgrade readiness, and provisioning new node groups in response to capacity events. Jentic exposes EKS operations as discoverable tools so an agent can search by intent and execute the matching operation without parsing AWS docs. This shortens incident response and capacity planning loops from manual console work to seconds.

Example prompt: Search Jentic for create cluster, load the EKS createCluster schema, and execute it for a new staging cluster in us-east-2.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | /clusters | Create a new EKS cluster |
| GET | /clusters | List EKS clusters in the account |
| DELETE | /clusters/{name} | Delete an EKS cluster |
| POST | /clusters/{name}/node-groups | Create a managed node group |
| POST | /clusters/{name}/fargate-profiles | Create a Fargate profile |
| POST | /clusters/{name}/addons | Install an EKS add-on |
| POST | /clusters/{name}/identity-provider-configs/associate | Associate an OIDC identity provider |

## Key resources

- **Clusters** — Create, describe, update, and delete EKS clusters and their control-plane configuration
- **Node Groups** — Manage managed node groups, including scaling, AMI updates, and instance configuration
- **Fargate Profiles** — Attach and detach Fargate profiles for serverless pod execution
- **Add-ons** — Install and update EKS-managed add-ons such as VPC CNI, CoreDNS, and kube-proxy
- **Identity Provider Configs** — Associate OIDC identity providers for Kubernetes authentication

## Why Jentic

- **Setup:** Wiring Amazon EKS by hand means building SigV4 request signing, resolving the regional eks.{region}.amazonaws.com host, and handling AWS throttling and retries yourself. Through Jentic you install once, import EKS from the API Directory, store the AWS access keys once, and your agent calls it.
- **Permission scoping:** EKS puts the cluster name in the URL path (/clusters/{name}/...), so a rule can pin your agent to one cluster: it can manage node groups, Fargate profiles, and add-ons for that cluster and nothing else. You choose the operations it may call, so cluster deletion is not included unless you add it.
- **Credential handling:** Your AWS access keys for EKS are stored once, encrypted, by your own Jentic One instance and injected at execution time when the request is signed. They never enter the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'provision a Kubernetes cluster' or 'add a managed node group', and Jentic returns the matching EKS operation with its input schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **Amazon EC2** — EC2 provides the underlying compute that EKS managed node groups run on
- **Amazon EMR on EKS** — Runs Spark and other big-data frameworks on existing EKS clusters
- **AWS Lambda** — Serverless functions for short-lived event-driven workloads

## FAQ

### Why is there no official OpenAPI spec for Amazon Elastic Kubernetes Service?

AWS does not publish an OpenAPI specification. Jentic generates and maintains this spec so that AI agents and developers can call Amazon Elastic Kubernetes Service via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer.

### What authentication does the Amazon EKS API use?

EKS uses AWS Signature Version 4 request signing (the spec marks the scheme as hmac/apiKey). Through Jentic, AWS access keys are stored encrypted in the vault and signing happens server-side, so raw credentials never enter the agent context.

### Can I create managed node groups with the EKS API?

Yes. POST /clusters/{name}/node-groups creates a managed node group with instance type, scaling configuration, AMI type, and Kubernetes labels. The cluster must already exist before you call this endpoint.

### What are the rate limits for the Amazon EKS API?

AWS applies per-account, per-region API throttling on EKS. Mutating calls (create cluster, create node group) are limited more aggressively than describe calls. Use exponential backoff on ThrottlingException responses; AWS does not publish exact TPS numbers in the spec.

### How do I install an EKS add-on through Jentic?

Search Jentic for install eks addon, load the schema for POST /clusters/{name}/addons, and execute it with the cluster name, addonName (for example vpc-cni), and addonVersion. Jentic returns the createdAt timestamp and add-on ARN once the call succeeds.

### Does the EKS API let me run kubectl commands?

No. The EKS API manages the cluster control plane, node groups, and add-ons. Kubernetes object operations (kubectl apply, get pods) go directly to the cluster's Kubernetes API server using the kubeconfig you obtain from describeCluster.

### Can I limit what my agent is allowed to do with the Amazon EKS API?

Yes. Because you run Jentic One yourself, your own rules decide which EKS operations and credentials the agent may use. Since EKS puts the cluster name in the URL path (/clusters/{name}/...), you can pin the agent to a single cluster so it manages that cluster's node groups, Fargate profiles, and add-ons and nothing else. You also choose the exact operations it may call, so destructive actions like deleting a cluster are excluded unless you explicitly allow them.