AI L1 Support Agent

This project demonstrates an AI agent for L1 support of Airflow failures, integrating with GitHub Issues (as a stand-in ticket system), RAG retrieval, and automated task reruns.

Structure

• airflow_dags/: DAG definitions with failure callbacks to GitHub Issues
• ingestion/: Scripts to build a FAISS-based RAG index from historical errors
• agent/: FastAPI service that ingests GitHub Issues, retrieves remediation, and triggers Airflow tasks
• .env.example: Sample environment variable file
• docker-compose.yml: Compose file to run the Agent
• requirements.txt: Global Python dependencies (optional)

Setup

1. Clone & configure

   git clone https://github.com/your-org/ai-l1-support.git
   cd ai-l1-support
   cp .env.example .env
   # Edit .env with your GH, Airflow & (optional) OpenAI/HF creds