PL-300 - Deploy and Maintain Assets (18% of the exam) - Section 4.2

Manage semantic models including scheduled refresh, endorsement, and query caching.

Configure scheduled refresh and incremental refresh on semantic models to keep data current, and apply dataset endorsement to mark certified or promoted content for consumers. Understand how query caching and the XMLA endpoint affect performance and external tool connectivity.

scheduled refreshincremental refreshdataset endorsementquery cachingXMLA endpoint

Practice question for this objective

Free sampleDeploy and Maintain Assetsmedium

A semantic model uses DirectQuery against an Azure SQL Database source. Scheduled refresh in the Power BI service refreshes the underlying source data for this DirectQuery model. Is this statement correct?

  • AYes
  • BNo Correct
DirectQuery and live connection models do not rely on scheduled refresh to update source data. In DirectQuery or live connection scenarios, most source data is not imported, so the model does not rely on scheduled refresh. Queries go to the source at view time instead.

Why A is wrong: Scheduled refresh applies to imported semantic models. DirectQuery only refreshes tile caches and metadata, not the underlying source rows.

Why B is correct: Correct. In DirectQuery or live connection scenarios, most source data is not imported, so the model does not rely on scheduled refresh.

See more PL-300 practice questions, answers explained.

Exam traps in Deploy and Maintain Assets

Answers that look right on this material and are not. Each one is a distractor from a different question in the PL-300 bank for this domain.

  • Apply the Certified endorsement on the semantic model settings.

    Why it is wrong: Certification can only be applied by authorised reviewers defined by the Power BI administrator.

  • Semantic models, dashboards, workspaces, and apps can be endorsed.

    Why it is wrong: Dashboards and workspaces themselves are not endorseable; dataflows (legacy) and reports are.

  • The consecutive failure threshold can be raised to 10 in semantic model settings to absorb flaky sources.

    Why it is wrong: The documentation states the four-failure threshold cannot be changed.

Examworthy is not affiliated with or endorsed by Microsoft. Original, blueprint-aligned practice material only.