Amazon ECS Fully Managed MCP Server

What is this?

The Amazon ECS Fully Managed MCP Server is a fully managed, AWS-native implementation of the Model–Chain–Protocol (MCP) that delivers secure, real-time context and lifecycle management for Amazon ECS clusters. It abstracts operational complexity by providing a stable, AWS-supported endpoint with automated scaling, patching, and security, so developers and AI agents can instantly query tasks, services, events, logs, and metrics without managing infrastructure.

This server implements the MCP 1.0 wire protocol using JSON-RPC over HTTP/2 and gRPC streaming, with mutual TLS authentication and fine-grained AWS IAM authorization. You can subscribe to cluster events, perform lifecycle operations, and capture audit logs in CloudTrail, all through simple API calls that integrate seamlessly with your existing AI and DevOps toolchain.

Quick Start

Install the server using npm:

npm install @modelcontextprotocol/server-aws-ecs-mcp-server

Then add it to your MCP client configuration:

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

Key Features

Feature 1: Zero-maintenance hosting with automated updates, high availability, and built-in scaling, so you never manage servers or patches.

Feature 2: Real-time ECS cluster context via gRPC streaming and JSON-RPC APIs, delivering task statuses, service health, events, and metrics with sub-second latency.

Feature 3: Enterprise-grade security and compliance powered by AWS IAM for fine-grained access control and CloudTrail for audit logging.

Example Usage

Let’s say you want to fetch the current status of your production cluster to power a dashboard or chat bot.

// Example code
const result = await client.callTool({
  name: "getClusterStatus",
  arguments: {
    cluster: "prod-cluster"
  }
});

This code calls the getClusterStatus tool on the MCP server and returns active services, running tasks, pending tasks, and overall cluster health in a simple JSON response.

Configuration

The server accepts the following environment variables:

API_KEY – Your API key or IAM credentials used for mTLS authentication and authorization via AWS IAM.

SETTING_1 (optional) – When set to true, enables debug logging output for troubleshooting.

Available Tools/Resources

ClusterResource: Perform list, describe, and update operations on ECS clusters.

ServiceResource: Create, update, and delete ECS services programmatically.

Who Should Use This?

This server is perfect for:

Use case 1: DevOps engineers who need real-time visibility and control over ECS clusters without managing infrastructure.

Use case 2: AI developers building context-aware chat bots or code generators that require up-to-date ECS data.

Use case 3: Monitoring teams integrating ECS metrics and events into observability platforms like Datadog or New Relic.

Conclusion

The Amazon ECS Fully Managed MCP Server gives you a turnkey way to bring secure, real-time ECS cluster context into your AI assistants, dashboards, and automation scripts. With zero maintenance, enterprise-grade security, and seamless integration, you can accelerate DevOps workflows and focus on building intelligent experiences.

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