autogen-issue-resolver

AutoGen Issue Resolver Prototype

This is a simple prototyp using AutoGen Multi-Agent framework to resolve issues in a distributed system.

How to run

To use the AutoGen Assistant Agents on OpenAI GPT-4o you need to have an OpenAI account and API key and need to set it in the file agts/agent.py.

This prototype needs docker and docker-compose to run. To start the system, run the following command in the root directory of the project:

docker compose up

This will start the system with 5 containers:

1. Service 1 (Python Flask App, svc1) - generating a random number between 1 and 500 README
2. Service 2 (Python Flask App, svc2) - requesting the random number from Service 1 and adding 50 README
3. Service 3 (Python Flask App, svc3) - requesting the calculated number from Service 2 README
4. A 4th Python Flask App (agts) running 4 AutoGen Assistant Agents (1 Issue Resolver & 3 Service Observer) - investigating issues in the system README
5. A VueJS frontend (frontend) to visualize the system state and AutoGen agent results README

After the system is started, you can access the frontend at http://localhost:8080.

This image displays the flow of the system: