Mix and Match: Serving an ADK Agent to AWS and Azure
This article provides a step by step look at running a Google ADK agent on Cloud Run, and serving it over the A2A protocol to callers that are not ADK. 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 credentials between them. Google…
This project aims to deploy three agents on Google Cloud, AWS, and Azure, allowing them to communicate using the A2A protocol without any stored credentials between them. The Google side of the project involves running an ADK agent on Cloud Run, which is then served over the A2A protocol to callers that are not ADK. The Google ADK framework is open source and model agnostic, enabling it to run Gemini through Vertex AI or an API key.
The A2A protocol allows agents built by different teams on different frameworks to call each other. The server is deployed using Google's code, and the agent card is created, which includes the URL to the server. The card only advertises a public HTTPS endpoint with unroutable plaintext. After deployment, the server's card is fetched, and the interface URL is taken from the card.
The same reply is delivered twice because ADK attaches the reply as a task artifact, and also leaves a copy in task history. When clients read artifacts, the content is empty, resulting in a successful call with no content. To fix this, every carrier the spec allows should be read, and ADK's reply now appears twice.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
