AWS re:Invent 2025 - Build AI-powered developer experiences with MCP on ECS and EKS (CNS358)

Building AI-Powered Developer Experiences with MCP on ECS and EKS

Overview

This presentation discusses how Amazon Web Services (AWS) is enabling AI-powered developer experiences through the use of the Model Context Protocol (MCP) on their container services, Elastic Container Service (ECS) and Elastic Kubernetes Service (EKS). The speakers provide a detailed introduction to MCP, explain the limitations of traditional AI assistants, and demonstrate how MCP can be leveraged to enhance developer productivity and troubleshooting capabilities.

The Evolution of AI Assistants and the Need for MCP

• Traditional AI assistants and large language models (LLMs) face limitations in understanding context, accessing real-time data, and taking actions beyond providing information.
• Approaches like knowledge bases (Rags) and dynamic tool integrations have helped, but still have limitations in scalability and flexibility.
• MCP is an open standard that enables standardized integration between AI applications and external data sources and tools, addressing the shortcomings of previous approaches.

Key Components of MCP

• MCP Client: The AI application or agent that uses the MCP protocol to interact with external resources.
• MCP Server: Consists of three main components:
  1. Tools: Model-controlled functions that the LLM can leverage, such as price lookups, web searches, or database queries.
  2. Resources: Application-controlled data sources, like files, database records, or user history.
  3. Prompts: Predefined prompts that the user can discover and use to interact with the MCP server.

Benefits of MCP for Container Services

• AI is transforming how developers work, build, and observe their code, creating a need for AI-powered developer experiences.
• Traditional AI assistants lack real-time awareness of new features and optimal usage patterns for container services like ECS and EKS.
• MCP servers can provide up-to-date information, guidance, and troubleshooting capabilities tailored to the container services, enhancing developer productivity.

AWS Hosted MCP Servers for ECS and EKS

• AWS has launched fully managed, remotely hosted MCP servers for both ECS and EKS, available in all commercial AWS regions.
• Key capabilities of the EKS MCP server:
  • Cluster management: Tools for creating, deleting, and managing EKS clusters.
  • Kubernetes resource management: Tools for managing Kubernetes resources like pods and services.
  • Documentation and troubleshooting: Tools for searching documentation and accessing a knowledge base of EKS troubleshooting guidance.
• Key capabilities of the ECS MCP server:
  • Operational tools: Tools for monitoring deployments, investigating container health, and troubleshooting issues.
  • Resource management: Tools for managing ECS tasks and task definitions.
  • Troubleshooting tools: Integrating with Amazon CodeGuru to provide AI-powered troubleshooting recommendations.

Integrating with Hosted MCP Servers

• Prerequisites for using the hosted MCP servers:
  • Configuring IAM permissions to allow the MCP client to interact with the MCP server.
  • Setting up the MCP client (e.g., Kiro) to connect to the hosted MCP server endpoint.
• The hosted MCP servers use a proxy to handle the AWS Signature Version 4 (SigV4) authentication required to access AWS services.

Troubleshooting Integration in the AWS Console

• The ECS and EKS consoles now feature an "Inspect with Amazon CodeGuru" button that integrates the hosted MCP servers to provide AI-powered troubleshooting recommendations.
• The MCP server tools are used to gather relevant information, diagnose issues, and provide step-by-step guidance to resolve problems with ECS and EKS resources.

Key Takeaways

• MCP provides a standardized way for AI applications to interact with external data sources and tools, addressing the limitations of traditional AI assistants.
• AWS has launched hosted MCP servers for ECS and EKS, offering a range of tools and capabilities to enhance developer productivity and troubleshooting.
• The integration of MCP servers with the AWS console enables AI-powered troubleshooting and guidance, further improving the developer experience.
• Adopting MCP can help organizations leverage the power of AI to streamline container-based application development and operations.