Health MCP Server

This project provides an MCP server for accessing Azure Cloud Health information using custom health tools.

Prerequisites

• Python 3.8 or higher
• pip (Python package manager)
• Azure credentials set up for DefaultAzureCredential (e.g., via az login or environment variables)
• The following Python packages:
  • requests
  • azure-identity
  • mcp (ensure this is installed or available in your environment)

Installation (from PyPI)

To install the package from PyPI:

pip install health-mcp-server

This will install all required dependencies.

Running the MCP Server (after installation)

After installing, you can run the MCP server with:

python -m health_tools.HealthToolsMcpServer

Or, if you have set up a console script entry point, simply:

health-mcp-server

Note: Ensure your Azure credentials are set up (e.g., run az login or set environment variables for a service principal) before starting the server.

Installation (from source)

1. Clone this repository.

2. Install dependencies:

   pip install requests azure-identity
   

   If mcp is not available on PyPI, ensure it is installed or accessible in your environment.

Running the MCP Server

From the root directory, run:

mcp run src/HealthToolsMcpServer.py

The server will start and register the health tools.

Use VS Code

Add the MCP server to VS code.

Usage

The server exposes two main tools:

• GetEntityHealthPy: Get the health of an entity.
• GetHealthModelPy: Get the health model.

Refer to the code in src/HealthToolsMcpServer.py for details on tool parameters.

Prompts

Typical prompts:

"Give me the health of the health model {health_model_name} in subscription {subscription_id} and resource group {resource_group}".

Once you have setup the context - health model, subscription id and resource group - you can query freely. For example:

• "Summarize the health for my health model"
• "Do you see any patterns in the health model"
• "Give me more details on the health of {entity_name}"

Logging

Logs are written to c:\temp\McpServerPy.txt.

Troubleshooting

• Ensure your Azure credentials are correctly configured.
• Check the log file for error details if something goes wrong.