ramvlt/intellijidea-mcp-databricks

Databricks MCP Server

A Model Context Protocol (MCP) server that provides integration with Databricks, enabling AI assistants to interact with your Databricks workspace.

Features

• SQL Execution: Run SQL queries on Databricks clusters using Databricks Connect
• Workspace Management: List and read notebooks
• Cluster Information: Get details about clusters
• Jobs Management: List, monitor, and trigger job runs
• DBFS Operations: List and read files from Databricks File System

Prerequisites

• Node.js 18 or higher
• A Databricks workspace
• Databricks personal access token
• A running Databricks cluster (for SQL execution)

Installation

1. Clone or navigate to this repository:

cd /path/to/databricks-mcp-server

2. Install dependencies:

npm install

3. Build the server:

npm run build

Configuration

1. Copy the example environment file:

cp .env.example .env

2. Edit .env and configure your Databricks credentials:

DATABRICKS_HOST=https://your-workspace.cloud.databricks.com
DATABRICKS_TOKEN=your-databricks-token
DATABRICKS_CLUSTER_ID=your-cluster-id

Getting Databricks Credentials

1. DATABRICKS_HOST: Your Databricks workspace URL (e.g., https://dbc-12345678-9abc.cloud.databricks.com or https://adb-xxxxx.azuredatabricks.net)
2. DATABRICKS_TOKEN: Generate a personal access token:
   • Go to User Settings > Developer > Access Tokens
   • Click "Generate New Token"
   • Copy the token value
3. DATABRICKS_CLUSTER_ID: (Required for SQL execution and cluster operations)
   • Go to Compute in the sidebar
   • Click on your cluster name
   • Copy the Cluster ID from the URL or cluster details (e.g., 1234-567890-abc123)
   • Important: The cluster must be running or set to auto-start

Usage with Claude Desktop

Add this configuration to your Claude Desktop config file:

MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "databricks": {
      "command": "node",
      "args": ["/absolute/path/to/databricks-mcp-server/dist/index.js"],
      "env": {
        "DATABRICKS_HOST": "https://your-workspace.cloud.databricks.com",
        "DATABRICKS_TOKEN": "your-databricks-token",
        "DATABRICKS_CLUSTER_ID": "your-cluster-id"
      }
    }
  }
}

Restart Claude Desktop after updating the configuration.

Available Tools

execute_sql

Execute SQL queries on your Databricks cluster using Spark SQL.

Parameters:
- query (required): SQL query to execute
- max_rows (optional): Maximum rows to return (default: 100)

Note: Requires a running cluster. Queries are executed using Spark SQL.

list_notebooks

List notebooks in a workspace path.

Parameters:
- path (optional): Workspace path (default: "/")

get_notebook_content

Get the content of a notebook.

Parameters:
- path (required): Path to the notebook

list_clusters

List all clusters in the workspace.

get_cluster_info

Get detailed information about a specific cluster.

Parameters:
- cluster_id (required): The cluster ID

list_jobs

List all jobs in the workspace.

Parameters:
- limit (optional): Maximum jobs to return (default: 25)

get_job_runs

Get run history for a job.

Parameters:
- job_id (required): The job ID
- limit (optional): Maximum runs to return (default: 25)

run_job

Trigger a job run.

Parameters:
- job_id (required): The job ID to run

list_dbfs

List files in DBFS.

Parameters:
- path (optional): DBFS path (default: "/")

read_dbfs_file

Read content from a DBFS file.

Parameters:
- path (required): DBFS path to the file

Development

Run in development mode:

npm run dev

Build:

npm run build

Security Notes

• Never commit your .env file or expose your Databricks token
• Use environment variables or secure secret management for production
• Grant minimal necessary permissions to the Databricks token
• Consider using service principals instead of personal access tokens for production use

Troubleshooting

"DATABRICKS_HOST and DATABRICKS_TOKEN must be set"

Make sure your environment variables are properly configured in the Claude Desktop config or .env file.

SQL execution errors

• Verify that DATABRICKS_CLUSTER_ID is correctly set and points to a valid cluster
• Ensure the cluster is running (not terminated or stopped)
• Check that your token has permissions to execute commands on the cluster

Connection timeout

• Check that your Databricks workspace is accessible and the token is valid
• Verify the cluster is in "Running" state
• Ensure your network allows connections to Databricks

Command execution takes too long

• Cluster may be starting up (this can take several minutes)
• Consider using a cluster with auto-termination disabled for faster response times

License

ISC