Databricks Data Engineer Associate Question 519
Single answerA data engineering team is setting up multiple Databricks workspaces for their organization, each in a different region. They plan to use Unity Catalog for centralized governance and data sharing. As a best practice, how should they configure their metastores to optimize performance and minimize latency?
- A
Assign a single metastore in one region and connect all workspaces to it.
- B
Assign a separate metastore for each workspace in its respective region.
- C
Assign one metastore per region and colocate it with the workspaces in that region.
- D
Use a single global metastore for all workspaces and enable caching for performance.
Show answer and explanation
Correct answer: C
Explanation
Colocating metastores with the workspaces in the same region is a best practice because it optimizes performance by reducing latency and ensures efficient operation while using Unity Catalog for centralized governance. This approach balances performance and governance requirements effectively.
- A. Incorrect.
This option does not follow best practices because having a single metastore in one region can result in high latency for workspaces located in other regions.
- B. Incorrect.
This approach is not ideal for Unity Catalog as it fragments governance and data sharing, which Unity Catalog is designed to centralize.
- C. Correct.
This is the best practice because colocating metastores with workspaces in the same region reduces latency and ensures better performance while maintaining centralized governance using Unity Catalog.
- D. Incorrect.
While using a single global metastore might seem convenient, it can lead to latency issues for workspaces in regions far from the metastore's location, even with caching enabled.