ms-fabric-mcp-server

Get Spark driver logs for a notebook execution.

Retrieves the driver logs (stdout or stderr) from a completed notebook run. This is particularly useful for getting detailed error messages and Python tracebacks when a notebook fails.

Important Notes:

• Python exceptions and tracebacks appear in stdout, not stderr

• stderr contains Spark/system logs (typically larger)

• For failed notebooks, check stdout first for the Python error

• Look for "Error", "Exception", "Traceback" in the output

Use this tool when:

• A notebook execution failed and you need to see the Python error

• You want to debug notebook issues by examining driver logs

• You need to analyze Spark driver behavior (stderr)

Parameters: workspace_name: The display name of the workspace containing the notebook. notebook_name: Name of the notebook. job_instance_id: The job instance ID from execute_notebook or run_on_demand_job result. log_type: Type of log to retrieve - "stdout" (default) or "stderr". Use "stdout" for Python errors and print statements. Use "stderr" for Spark/system logs. max_lines: Maximum number of lines to return (default: 500, None for all). Returns the last N lines (most recent, where errors typically are).

Returns: Dictionary with: - status: "success" or "error" - message: Description of the result - log_type: Type of log retrieved - log_content: The actual log content as a string - log_size_bytes: Total size of the log file - truncated: Whether the log was truncated - spark_application_id: The Spark application ID - livy_id: The Livy session ID

Example: ```python # Get Python error from a failed notebook result = get_notebook_driver_logs( workspace_name="Analytics", notebook_name="ETL_Pipeline", job_instance_id="12345678-1234-1234-1234-123456789abc", log_type="stdout" # Python errors are in stdout! )