Question
You plan to deploy Azure Databricks to support a machine learning application. Data engineers will mount an Azure Data Lake Storage account to the Databricks file system. Permissions to folders are granted directly to the data engineers.
You need to recommend a design for the planned Databrick deployment. The solution must meet the following requirements:
✑ Ensure that the data engineers can only access folders to which they have permissions.
✑ Minimize development effort.
✑ Minimize costs.
What should you include in the recommendation? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.
Explanation
Databricks SKU: Premium. Azure Data Lake Storage credential passthrough is available only on the Premium plan, so Premium is the lowest-cost SKU that meets the access requirement. Cluster configuration: Credential passthrough. With credential passthrough, each data engineer reaches the Data Lake Storage account with their own Microsoft Entra ID identity, so the folder permissions granted to them are enforced and no code or extra configuration is needed. A secret scope or a managed identity would give everyone the same identity and could not enforce per-user folder permissions. MLflow and Photon have nothing to do with access control. Credential passthrough is now a legacy feature that Databricks recommends replacing with Unity Catalog, but it is the right answer for the options given.
