Deploying Python MCP on AWS ECS Express

Next article in my cross cloud Agent series - this installment covers building a Python MCP environment and then deploying the MCP Server to an Amazon ECS Express endpoint. ECS Express is a serverless container running system on AWS that offers simplified deployment options but still allows full customization where needed. The entire solution was validated with Gemini CLI running as a MCP client. The code base is virtually unchanged across clouds and is a building block to validate the approach. The Cross Cloud Series is here: https://lnkd.in/eHYsbUhH The Medium Article is here: https://lnkd.in/eDHE4-2T dev.to is here: https://lnkd.in/eYZDmf77 #GDE #MCP #Python #GooogleCloud #AWScommunity #AWS #AWSCommunityBuilders

There are multiple tools for Infrastructure as Code (IaC) on AWS. My favourites are Terraform and the Serverless Application Model (SAM). For quick and simple use cases I find SAM works really well. Below is an example of using it along with GitHub Actions. SAM simplifies the config for the most common serverless components on AWS and allows you to embed Cloudformation for any other AWS resource types. It has a really nice CLI and allows local testing. To automate the deployment using SAM an approach involving Github Actions works well. This article from Wes Huber shows all of this and includes deploying to AWS. I would recommend using OIDC for the auth instead of an access key but this example will get you started. https://lnkd.in/ebVP-yqh #SAM #IaC #GithubActions

Next article in my cross cloud Agent series - this installment covers building ADK agents and then deploying the complete solution to an Amazon cluster running AKS. A whole cluster just for one Agent? Probably overkill- but I wanted to start with the most basic approach before tackling more complex scenarios with multiple pods, agents, gateways, and proxies. The code base is virtually unchanged across clouds and is a building block to validate the approach. The Cross Cloud Series is here: https://lnkd.in/eHYsbUhH The Medium Article is here: https://lnkd.in/evpaZ8HY dev.to is here: https://lnkd.in/eJ_icFWJ #GDE #MCP #Python #GooogleCloud #AWScommunity #AWS #AWSCommunityBuilders

I have been posting non-stop from Google Cloud Next- but also made the latest AWS community builders list!! My article covered a multi-agent pipeline with an image generation model- all deployed to AWS ECS Express. ECS express gives you a lot of the benefits of Kubernates but without the management complexity. Article is here: https://lnkd.in/er9ZFcSw #MCP #Python #GoogleCloud #AWScommunity #AWS #AWSCommunityBuilders

Continuing my cross cloud series - this installment covers building a Python and ADK Agent environment with the Google Agent Development Kit and then deploying a multi-agent solution to an Amazon Fargate endpoint. Fargate is a serverless container running system on AWS that offers more customization options that Lightsail. This ADK deployment features the Gemini Live model, Websockets, and deep Python integration. The code base is virtually unchanged across clouds and is a building block to validate the approach. The Cross Cloud Series is here: https://lnkd.in/eHYsbUhH The Medium Article is here: https://lnkd.in/eTpRHSjr dev.to is here: https://lnkd.in/eiUmFXh7 #GDE #ADK #GooogleCloud #AWScommunity #AWS #AWSCommunityBuilders

🚀 #PythonJourney | Day 161 — API Live on Google Cloud Run After a full day of debugging, my URL Shortener API is now running on Google Cloud Run in production. The journey: Docker misconfiguration → Port binding issues → Environment variable magic → SUCCESS. Key learnings: → Cloud Run expects PORT=8080 by default → ${PORT:-8080} in shell scripts is your friend → Sometimes the simplest fix is the last one you try → Persistence beats perfection The API is live. Tests pass. It's real now. #Backend #GCP #CloudRun #DevOps #Python #API

🚀 I have shared a new blog! In this, I’ve explained how a FastAPI application can be containerized with Docker and deployed on AWS EC2 in a simple, practical way. 💡 One key takeaway: Getting an application to run consistently across environments is where real learning happens. This project helped me understand how real-world systems are built, deployed, and maintained beyond just writing code. I’d love to hear your thoughts 👇 👉 What would you improve or do differently in this setup? 🔗 Blog link in comments #AWS #Docker #FastAPI #Python #Cloud #Tech#Blog

Multi-Region AWS Infrastructure (Terraform + Python Validation) Just wrapped up a multi-region AWS infrastructure project using Terraform. Built out a full stack including: - EC2 + RDS application layer - ALB + Route53 DNS - CloudFront CDN with HTTPS + WAF - Transit Gateway peering between regions (us-west-2 ↔ sa-east-1) One thing I focused on was not just deploying resources, but validating them. I wrote Python (boto3) scripts to verify: - CloudFront distribution - ACM certificate - Route53 alias - ALB connectivity The screenshot shows a full validation pass across the stack. This project pushed me to understand not just “how to build,” but how to debug and verify cloud systems end-to-end. Next step: expanding into failover and multi-region resiliency.

I finally understood how a backend actually works — not just ran the code. For the past few days, I was learning Docker, Flask, and Redis. Things kept breaking, and I had to restart multiple times. But finally it clicked. I built a simple URL shortener: 👉 Takes a long URL 👉 Generates a short code 👉 Stores it 👉 Redirects back when opened While fixing errors, I started understanding how each part connects. ⚙️ Tech I used: Flask (handles requests) Redis (stores key → URL) Docker (runs everything together) AWS EC2 (server) Still learning, but this felt like real progress. 💻 GitHub:https://lnkd.in/ga9DvTT6 #DevOps #Docker #AWS #Python #Learning

Next article in my cross cloud Agent series - this installment covers building ADK agents and then deploying the complete solution to an Amazon ECS cluster running ECS Express. A whole cluster just for an Agent? Probably overkill- but I wanted to start with the most basic approach before tackling more complex scenarios with multiple agents, gateways, and proxies. The code base is virtually unchanged across clouds and is a building block to validate the approach. The Cross Cloud Series is here: https://lnkd.in/eHYsbUhH The Medium Article is here: https://lnkd.in/ebjqBYfu dev.to is here: https://lnkd.in/er9ZFcSw #GDE #MCP #Python #GooogleCloud #AWScommunity #AWS #AWSCommunityBuilders