Databricks MCP Server

# Databricks APIs Documentation This directory contains documentation for calling Databricks APIs directly using the Databricks SDK, without wrapper utilities. ## Overview The Databricks SDK provides direct access to all Databricks APIs. This documentation shows you how to call these APIs directly in your FastAPI application. ## Authentication All APIs use the same authentication configured in your app: ```python from databricks.sdk import WorkspaceClient # Automatically uses your configured authentication client = WorkspaceClient() ``` ## API Categories ### [Databricks SDK](databricks_sdk.md) - Comprehensive SDK usage patterns - Authentication and configuration - Core APIs overview - FastAPI integration examples ### [Workspace APIs](workspace_apis.md) - File operations (upload, download, list) - Directory management - Notebook operations - Search and metadata ### [MLflow GenAI](mlflow_genai.md) - AI agent tracing and observability - Automated quality evaluation - Feedback and continuous improvement - Application lifecycle management ### [Model Serving](model_serving.md) - List and inspect serving endpoints - Query serving endpoints - Batch inference - Error handling and best practices ## Common Patterns ### Error Handling ```python from databricks.sdk.errors import DatabricksError try: result = client.workspace.get_status('/path/to/file') except DatabricksError as e: print(f"API Error: {e}") ``` ### Pagination ```python # Most list operations support pagination for cluster in client.clusters.list(): print(f"Cluster: {cluster.cluster_name}") ``` ### Async Operations ```python # Many operations are async - wait for completion cluster_id = client.clusters.create(...) client.clusters.wait_get_cluster_running(cluster_id) ``` ## SDK Documentation - [Databricks SDK Python Documentation](https://databricks-sdk-py.readthedocs.io/en/latest/) - [Databricks REST API Reference](https://docs.databricks.com/api/workspace/introduction) - [SDK Examples](https://github.com/databricks/databricks-sdk-py/tree/main/examples) ## Best Practices 1. **Reuse the client**: Create one `WorkspaceClient` instance per application 2. **Handle errors**: Always wrap API calls in try-catch blocks 3. **Use pagination**: Don't assume all results fit in one response 4. **Check permissions**: Verify user has required permissions for operations 5. **Log API calls**: Add logging for debugging and monitoring