Amazon EKS MCP Server

What is this?

The Amazon EKS MCP Server is a fully managed Multi-Channel Protocol (MCP) server for Amazon EKS clusters, packaged as @amazon-eks/mcp-server. It eliminates the need to install or maintain local AI agents by hosting and auto-scaling the service in your AWS account, delivering real-time cluster state, AWS best practices, and security controls through any MCP-compatible client.

Built on AWS Fargate, the server uses a microservices architecture—combining an MCP Gateway, Context Engine, Knowledge Graph, and on-demand Tools Manager—to continuously ingest EKS API data, CloudTrail events, and AWS documentation. Every action is secured by AWS IAM policies and audited via CloudTrail, while automatic updates and deep integration with tools like Claude Desktop and Cursor IDE ensure seamless AI-powered CI/CD, troubleshooting, and cluster configuration workflows.

Quick Start

Install the server using npm:

npm install @amazon-eks/mcp-server

Then add it to your MCP client configuration:

{
  "mcpServers": {
    "amazon-eks-mcp-server": {
      "command": "npx",
      "args": ["-y", "@amazon-eks/mcp-server"],
      "env": {
        "API_KEY": "your-api-key-here"
      }
    }
  }
}

Key Features

Cluster Context Awareness: Maintains up-to-date state on nodes, pods, policies, and services to power context-aware AI prompts and recommendations.

Secure by Design: Enforces AWS IAM policies and audits every action through CloudTrail for full compliance and traceability.

Automatic Maintenance: Delivers zero-downtime updates and automatic patching, eliminating manual upgrades and reducing operational overhead.

Example Usage

Here’s how you might use the server to optimize your cluster configuration in a development script:

// Example code
const result = await client.callTool({
  name: "optimize-cluster",
  arguments: {
    clusterName: "my-eks-cluster"
  }
});

In this example, the optimize-cluster tool analyzes your cluster topology and recommends cost and availability improvements, returning structured suggestions you can apply directly.

Configuration

The server accepts the following environment variables:

API_KEY – Your AWS credentials or API key used for SigV4 authentication and secure access.

SETTING_1 (optional) – An optional configuration setting such as CUSTOM_ENDPOINT to override the default server URL.

Available Tools/Resources

kubectl: Executes Kubernetes commands and retrieves cluster data.

helm: Manages Helm charts, performs diffs, and deploys applications.

Who Should Use This?

This server is perfect for:

DevOps Engineers: Automate cluster monitoring, scaling, and cost optimizations with AI-driven insights.

Developers: Integrate AI-assisted manifest validation and best practice recommendations into CI/CD pipelines.

SREs: Ensure secure, compliant operations and troubleshooting with audit-ready logs and contextual debugging.

Conclusion

The Amazon EKS MCP Server brings real-time context, AWS best practices, and enterprise-grade AI capabilities to your Kubernetes workflows. Get started today to streamline cluster management, CI/CD automation, and troubleshooting with a fully managed backend maintained by AWS.

Check out the GitHub repository for more information and to contribute.