MCP.so
Sign In

Azure Kusto MCP Server

@ashishroy077

About Azure Kusto MCP Server

An MCP Server for Azure Kusto

Basic information

Category

Cloud & Infrastructure

Runtime

python

Transports

stdio

Publisher

ashishroy077

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "Azure Kusto MCP": {
      "command": "python",
      "args": [
        "-m",
        "src.kusto_mcp.server"
      ],
      "env": {
        "AZURE_KUSTO_CLUSTER": "",
        "AZURE_KUSTO_DATABASE": ""
      }
    }
  }
}

Tools

5

Connect to a Kusto cluster and database

Check the current connection status

Run a KQL query

Execute a query and analyze the results

Get suggestions for query optimization

Overview

What is Azure Kusto MCP Server?

A Model Context Protocol (MCP) server that connects to Azure Kusto, enabling AI assistants to explore data schemas and execute KQL queries. It is designed for developers and data analysts using VS Code with GitHub Copilot.

How to use Azure Kusto MCP Server?

Run python setup-mcp.py for quick integration with VS Code, or manually install dependencies and create an mcp.json file in the .vscode directory. Start the server by running python -m src.kusto_mcp.server or through VS Code’s MCP command palette.

Key features of Azure Kusto MCP Server

  • Secure Azure Kusto integration with DefaultAzureCredential authentication.
  • Schema exploration via resource URIs (e.g., kusto/tables, kusto/schema/{table}).
  • Tools to connect, execute queries, analyze data, and optimize KQL.
  • Built-in prompts for common data analysis tasks.
  • VS Code integration with interactive cluster/database input prompts.

Use cases of Azure Kusto MCP Server

  • AI assistants browsing table schemas to answer schema-related questions.
  • Executing KQL queries and analyzing results directly from Copilot Chat.
  • Optimizing complex KQL queries through automated suggestions.
  • Performing time-series or correlation analysis on Kusto data.
  • Quality checks for null values and outliers in datasets.

FAQ from Azure Kusto MCP Server

How does authentication work?

The server uses Azure’s DefaultAzureCredential, supporting environment variables, Managed Identity, Azure CLI, PowerShell, and interactive browser fallback.

What are the runtime requirements?

Python 3.9+, an Azure Kusto cluster, and VS Code with GitHub Copilot or Copilot Chat extension.

Where are credentials stored?

Credentials are never stored by the server; they are handled by Azure’s credential chain.

Are queries read-only?

Yes, the server executes read-only KQL queries and provides analysis tools.

What transports are supported?

The server uses stdio transport, configured via a VS Code mcp.json file.

Comments

More Cloud & Infrastructure MCP servers