Exception raised in a notebook or Python script. SQL syntax error or missing table in a SQL task. Assertion or data quality check failed. Library import error or missing dependency. Insufficient permissions to access a data source

Open the failed run in the Databricks Jobs UI and click on the failed task. Expand the error output to see the full stack trace or error message. Identify whether the error is in user code (notebook/script), a SQL statement, or an infrastructure issue. Use the 'Repair Run' feature to retry only the failed task after fixing the issue. Check the cluster event log for infrastructure-level errors if no code error is visible

High severityjob

Power BI Refresh Error:
FAILED

What does this error mean?

A Databricks job run completed with an error. The task executed but encountered a failure — this is the most common terminal error state in Databricks Jobs.

Common causes

1Exception raised in a notebook or Python script
2SQL syntax error or missing table in a SQL task
3Assertion or data quality check failed
4Library import error or missing dependency
5Insufficient permissions to access a data source

How to fix it

1Open the failed run in the Databricks Jobs UI and click on the failed task
2Expand the error output to see the full stack trace or error message
3Identify whether the error is in user code (notebook/script), a SQL statement, or an infrastructure issue
4Use the 'Repair Run' feature to retry only the failed task after fixing the issue
5Check the cluster event log for infrastructure-level errors if no code error is visible

Frequently asked questions

Can I retry only the failed tasks?

Yes — Databricks supports 'Repair Run', which reruns only failed and skipped tasks without rerunning successful ones.

How is FAILED different from INTERNAL_ERROR?

FAILED means user code or a data issue caused the failure. INTERNAL_ERROR means the Databricks platform itself encountered an unexpected error.