Question

You manage a Fabric workspace where several notebooks run memory-intensive machine learning workloads. Data engineers report that jobs frequently start with default compute that is too small, and they want a reusable configuration that pins a specific Spark runtime version, custom library set, and a larger node size for these notebooks—without editing each notebook individually. What should you configure in the workspace settings?

Accepted Answer

Create a Fabric Environment that defines the Spark runtime, libraries, and compute pool, then set it as the workspace default

Answer

Create a custom Spark pool and attach it directly to each notebook cell using the %%configure magic command

Answer

Enable high concurrency mode in the workspace Spark settings so sessions share one large driver

Answer

Increase the capacity SKU for the entire workspace to guarantee larger node sizes for all items