What problem does it solve? Working with Microsoft Fabric Spark through raw REST APIs is error-prone: wrong payload shapes, missed LRO polling, ambiguous workspace/item IDs, and confusion between notebook jobs, Livy sessions, and Materialized Lake View operations. This Skill routes each request to the correct operational mode and executes the documented Fabric API calls against the live workspace. ## Core Features & Use Cases - Notebook authoring and execution: Write PySpark, Scala, SparkR, %%sql, and %%configure cells, save them via updateDefinition, and run named notebooks through the Jobs API. - Livy-session analytics: Create Lakehouse Livy sessions and run interactive ad-hoc PySpark calculations, cross-lakehouse joins, and Delta time-travel queries. - Spark failure triage: Diagnose failed, slow, or throttled notebook, pipeline, and Livy runs using Spark Advisor, monitoring APIs, logs, and resource-usage metrics. - Materialized Lake View lifecycle: Author MLV definitions, review queries for incremental-refresh readiness, create refresh schedules, trigger on-demand refreshes, and classify refresh failures. - Use Case: A user asks "My notebook ETL_Daily failed with an OOM in workspace Production — find out why." The Skill resolves the workspace and notebook, queries Spark Advisor and stage metrics, identifies data skew, and reports the root cause with fix recommendations. ## Quick Start Ask the assistant to write a notebook cell, run a notebook by name, create a Livy session, or diagnose a failed Spark run in your Fabric workspace, naming the workspace and item involved.
