Bret and Nirmal are joined by Rajdeep Saha from AWS to discuss AI agents and the MCP (Model Context Protocol) standard. In this episode, they explore how AI can manage infrastructure with ease and demonstrate creating Kubernetes clusters with a single prompt.
😇 My new GitHub Security workshop has launched! A free 2-hour workshop with hands-on labs to harden your repos and your workflows from common supply chain attacks. I'll cover how attackers are getting in, and then we'll lock down a sample repo so you know what needs to be done to protect your code. You'll leave with a deep understanding of risks and mitigations as well as a list of helpful tools to keep your repos safe, including my new "gasa" tool for scanning your repos and orgs.
We demonstrate AWS's new MCP server for EKS, showing how AI agents can create Kubernetes clusters and deploy applications with simple English prompts. Using Claude Sonnet 4 through the Cline VS Code extension, Raj creates an entire EKS cluster with CloudFormation, deploys a workload, and demonstrates how the AI automatically troubleshoots and fixes deployment issues - all without touching kubectl or writing YAML manually.
We explore the Model Context Protocol (MCP) standard and how it enables AI agents to dynamically discover and use tools, from AWS services to cost calculators. We discuss the security implications of giving AI write access to AWS accounts, the challenges of context window management, and why local LLMs struggle with complex agentic workflows compared to foundational models like Claude.
We cover emerging standards, particularly Google's Agent-to-Agent (A2A) protocol that was recently donated to CNCF. This complements MCP by standardizing how agents communicate with each other, potentially leading to agent marketplaces where specialized agents can collaborate on complex tasks.
We stress the need for proper security guardrails, best practices, and organizational readiness. As Raj puts it, "Gen AI is an awesome white horse, but you still need Gandalf to ride it."
Looking ahead, we predict rapid evolution in agentic frameworks, multi-agent support, and standardized authentication. For DevOps professionals, the bar is rising - tasks that once took a week may soon be expected in two days with AI assistance. It is becoming crucial to understand these tools before they reach mainstream.
Check out the video podcast version here: https://youtu.be/dAkNUoZ_xtw
Creators & Guests
- (00:00) - Introduction
- (05:28) - Why Use Cline with VS Code?
- (13:39) - Security and Best Practices
- (15:24) - MCP "Marketplaces"
- (23:22) - Understanding MCP Endpoint Calls
- (24:51) - AWS Account Communication and Tool Access
- (25:18) - Cost Estimation with Serverless MCP
- (26:21) - Agentic Systems and Iterative Feedback Loops
- (28:41) - Deploying and Debugging with EKS MCP Server
- (31:19) - Challenges with LLMs in Troubleshooting
- (38:10) - Agent to Agent Communication (A2A)
- (45:10) - Anticipating Future Developments
- (45:47) - Practical Advice for Cloud Engineers
- (49:40) - Amazon Q
- (56:45) - Conclusion and Final Thoughts
You can also support this podcast by subscribing to my YouTube channel and my weekly newsletter at bret.news!
Grab the best coupons for my Docker and Kubernetes courses on Udemy.
Join my cloud native DevOps community on Discord.
Grab some merch at Bret's Loot Box
Homepage bretfisher.com
★ Support this podcast on Patreon ★
