Question

A scheduled Fabric notebook that loads data into a lakehouse Delta table has started failing. When you open the notebook run in Monitoring hub, you see the job failed but the summary only shows a generic 'Notebook execution failed' message. You need to pinpoint the specific line of code and exception that caused the failure with the least effort. What should you do?

Accepted Answer

Open the failed run's detail view and inspect the cell-level output and the Spark driver log / stack trace for the exception raised by the failing cell

Answer

Re-run the entire notebook interactively and wait for it to fail again to observe the error

Answer

Enable Spark autoscale on the workspace pool and rerun the scheduled job

Answer

Review the Capacity Metrics app to check whether the notebook was throttled by the capacity