Mix and Match: Serving a Bedrock Agent to Google and Azure
This article provides a step by step guide for running a Strands agent on Amazon Bedrock AgentCore Runtime, and serving it over the A2A protocol to callers on Google Cloud and Azure. The code is here: github.com/xbill9/multicloud-a2a-subagent What is this project trying to Do? This project aims to answer one research brief with three agents on three clouds, over one protocol, with no stored…
This project seeks to serve three agents – one on Amazon Bedrock, one on Google Cloud, and one on Azure – through a single protocol, without sharing any stored credentials between them. On the AWS side, an ADK agent runs on Cloud Run, an Agent Framework agent runs on Container Apps, and a Strands agent runs on AgentCore.
The Strands agent, which is open-source, requires only a few lines of code to run a model, system prompt, and tools. However, it does not have built-in A2A (Agent2Agent) server integration. The code for the Strands agent is then wrapped with the a2a-sdk reference routes to serve the protocol's own implementation. The runtime for the Strands agent on AWS must be ARM64, and the container must listen on port 9000, not 8080 as in the case of Cloud Run or Container Apps.
To deploy the Strands agent on AWS, a single command builds the ARM64 image, pushes it to Amazon ECR, creates the runtime and IAM roles, and then the endpoint can be obtained. The deployment script also includes verification of the installation, which includes two curl tests that should fail without proper authentication and a positive control that verifies the coordinator on Google Cloud can mint a token, ensuring the cross-cloud integration works correctly.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
