Supercharge Kubernetes with Amazon EKS MCP Server & AI Automation

Managing Amazon EKS clusters can be complex, especially for API and backend developers who want to move fast without wrestling with YAML files or kubectl commands. What if you could use AI to automate EKS tasks by simply chatting in your IDE?

In this step-by-step guide, you'll learn how to use the Amazon EKS MCP Server—an open-source tool that empowers AI assistants (like Cline) to create EKS clusters, deploy workloads, and troubleshoot Kubernetes, all through natural language. We'll walk through installation in VS Code, deploying NGINX, and best practices for troubleshooting, making cloud-native operations accessible and efficient for modern API teams.

If you need an all-in-one API platform that generates beautiful API Documentation, boosts team productivity, and replaces Postman at a much more affordable price, Apidog is a top choice for API-focused teams.

What is Amazon EKS MCP Server?

Amazon EKS MCP Server is an open-source bridge between AI code assistants and Amazon Elastic Kubernetes Service (EKS). It uses the Model Context Protocol (MCP) to enable tools like Cline, Amazon Q, or Cursor to manage EKS clusters via natural language prompts.

Key Features:

• Cluster Management: Create and manage EKS clusters with auto-configured VPCs and IAM roles.
• Resource Automation: Deploy, update, and monitor Kubernetes resources (pods, services, deployments) using YAML or conversational English.
• Troubleshooting: Access logs, CloudWatch metrics, and events for real-time debugging.
• Secure by Default: Runs in read-only mode unless explicitly allowed for write operations.

The EKS MCP Server is like adding an AI-powered DevOps expert to your workflow—speeding up cluster ops and removing manual toil.

Why Use Amazon EKS MCP Server?

Amazon EKS MCP Server streamlines Kubernetes management for API and backend engineers, offering:

• Simplicity: Manage clusters and workloads without deep Kubernetes or eksctl expertise.
• Automation: Let AI assistants automate repetitive tasks, freeing you to focus on building and testing APIs.
• Team Productivity: Standardize EKS operations across teams, reducing errors and onboarding time.
• AI Integration: Use LLMs for natural language-driven DevOps, ideal for fast-moving API teams.

Example: I spun up a new EKS cluster and deployed NGINX in under 15 minutes—no manual YAML!

Step-by-Step: Install & Use Amazon EKS MCP Server with Cline in VS Code

Below, you'll find a practical walkthrough designed for technical users. We'll focus on using Cline in VS Code for the smoothest experience, but also cover manual setup.

Prerequisites

Before starting, ensure you have:

• Python 3.10+ (python3 --version)
• AWS CLI (configured with EKS permissions; check with aws sts get-caller-identity)
• VS Code (latest version)
• Cline Extension (AI assistant)
• AWS Account (with IAM roles for EKS, VPC, CloudFormation)
• kubectl (for direct pod management)
• uv (Python package manager: pip install uv)

Manual Installation (Quick Overview)

For manual setup:

git clone https://github.com/awslabs/mcp.git
cd mcp/src/eks-mcp-server
uv pip install awslabs.eks-mcp-server
uv run main.py

• Configure AWS credentials in ~/.aws/credentials.
• Start the server and ensure permissions are set.

For most users, the Cline extension in VS Code offers a more seamless experience.

Installing Cline in VS Code

1. Open VS Code, navigate to Extensions (Ctrl+Shift+X), and search for Cline.
2. Click Install.
3. Open the Cline chat panel (Ctrl+Shift+P → "Cline: Open Chat").
4. Configure an LLM provider (e.g., AWS Bedrock or OpenAI) via Cline settings (gear icon > API Provider).

Tip: AWS Bedrock offers tight AWS integration. Setup takes just a couple of minutes!

Setting Up Amazon EKS MCP Server via Cline

1. Open Cline’s chat panel.

2. Prompt Cline:

   • Basic:
     Install eks-mcp-server from awslabs
   • Detailed:
     Install the MCP server named awslabs.eks-mcp-server for Cline. Use uvx to run the server and update mcp settings.

3. Approve Cline's actions (e.g., running uvx awslabs.eks-mcp-server@latest).

Cline updates your ~/.aws/amazonq/mcp.json for tool configuration—no manual YAML editing required.

4. Verify Installation:
   In Cline’s chat, run /tools to list available EKS MCP tools like create_eks_cluster and apply_yaml.

Troubleshooting:
If you see errors (e.g., “docker not found”), add --allow-write to args and ensure eksctl is installed (brew install eksctl on macOS).

5. AWS Permissions:
   Run aws configure in your terminal and enter your AWS credentials if needed.

Pro Tip:
Install a web-search MCP server (like firecraw) from the Cline MCP marketplace to help Cline self-diagnose and fix setup issues if they arise.

Creating a New EKS Cluster with Cline

1. In Cline’s chat, prompt:

   • Basic:
     Help create a new EKS cluster
   • Custom:
     Help create a new EKS cluster named 'my-ai-cluster' in us-east-1 using the Amazon EKS MCP Server.

2. Cline uses the create_eks_cluster tool to automate VPC, subnet, and IAM setup.

3. Approve the action when prompted.

4. Monitor Progress:
   Cluster creation takes ~10 minutes. Check status with:

   eksctl get cluster --name my-ai-cluster --region us-east-1
   

   Output: my-ai-cluster is ACTIVE

Deploying NGINX with Cline

1. In Cline’s chat, request:

   Prepare a deployment file for NGINX. Create a comprehensive Kubernetes manifest that includes both a deployment and a service for NGINX using the Amazon EKS MCP Server.
   

2. Cline generates and applies the following manifest:

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: nginx-deployment
     namespace: default
   spec:
     replicas: 3
     selector:
       matchLabels:
         app: nginx
     template:
       metadata:
         labels:
           app: nginx
       spec:
         containers:
         - name: nginx
           image: nginx:latest
           ports:
           - containerPort: 80
   ---
   apiVersion: v1
   kind: Service
   metadata:
     name: nginx-service
     namespace: default
   spec:
     selector:
       app: nginx
     ports:
     - protocol: TCP
       port: 80
       targetPort: 80
     type: LoadBalancer
   

3. Approve the action.
4. Verify Deployment:

   kubectl get pods -A
   

   Example output:

   NAMESPACE  NAME                           READY  STATUS   RESTARTS  AGE
   default    nginx-deployment-abc123-xyz    1/1    Running  0         2m
   

Result: Your NGINX service is live in minutes, with zero manual YAML tweaking!

Troubleshooting Pod Issues with Cline

If your pods are not running (e.g., CrashLoopBackOff):

1. Check pod status:

   kubectl get pods -A
   

   Example:

   NAMESPACE  NAME                           READY  STATUS            RESTARTS  AGE
   default    nginx-deployment-abc123-xyz    0/1    CrashLoopBackOff  3         5m
   

2. Prompt Cline:

   Figure out the issue with my pods and fix it using the Amazon EKS MCP Server.
   

3. Cline uses tools like get_pod_logs and get_k8s_events to diagnose, then suggests and applies fixes (e.g., updating the image version).

4. Re-check status:

   kubectl get pods -A
   

   NAMESPACE  NAME                           READY  STATUS   RESTARTS  AGE
   default    nginx-deployment-abc123-xyz    1/1    Running  0         1m
   

AI-driven troubleshooting helps you quickly resolve deployment issues with minimal effort.

Common Issues & Troubleshooting Tips

• Cline Errors:
  Check mcp.json for correct AWS_PROFILE and --allow-write permissions.
• Permission Denied:
  Ensure IAM roles have EKS, VPC, and CloudFormation access.
• Port Conflicts:

  lsof -i :3000
  kill -9 [PID]
  

• Pod Failures:
  Use Cline’s search_eks_troubleshoot_guide tool or:

  kubectl logs [pod-name]
  

• Documentation & Help:
  Visit awslabs.github.io/mcp for detailed guides.

Customizing & Extending Amazon EKS MCP Server

Power users can enhance their setup by:

• Multi-Region Support: Add multiple AWS_REGION values in mcp.json.
• Custom Tools: Extend your MCP server with new tools from the MCP repo.
• Production Safety: Run in read-only mode (ALLOW_WRITE=false) to prevent accidental changes.
• IaC Integration: Combine with AWS CDK or Terraform MCP servers for advanced infrastructure as code workflows.

Example: I added support for us-west-2 in under 5 minutes for multi-region testing.

Why API Teams Love Amazon EKS MCP Server

With Amazon EKS MCP Server, Kubernetes management becomes conversational, fast, and reliable. It’s ideal for API developers and teams who want AI-powered automation without sacrificing control or security.

Compared to manual eksctl and kubectl workflows, this approach saves time, reduces errors, and lets your team focus on building great APIs.

If you’re looking to further streamline your API lifecycle—from testing to documentation—consider Apidog for beautiful API docs, integrated team workflows, and the productivity edge your team deserves. Apidog is trusted by API teams seeking a modern, cost-effective alternative to Postman.