MCP Airflow Database

A Model Context Protocol (MCP) server for interacting with Airflow databases.

Setup with Poetry

Prerequisites

• Python 3.8 or higher
• Poetry installed on your system

Installation

1. Clone this repository:

   git clone <your-repository-url>
   cd mcp-airflow-db
   

2. Install dependencies with Poetry:

   poetry install
   

3. Configure your environment: Create a .env file with your database connection string:

   DATABASE_URL=postgresql://airflow:airflow123@localhost:5432/airflow
   

Running the MCP Server

Run the server with Poetry:

poetry run python src/server.py

Or activate the Poetry environment first:

poetry shell
python src/server.py

Using with Smithery

This MCP can be used with Smithery directly as configured in the smithery.yaml file. Make sure to provide the DATABASE_URL configuration when starting the server.

Available Tools

• failed_runs: Query failed Airflow DAG runs within a specified time period.
• query: Execute SQL queries directly against the Airflow database.

License

[Your License]