Plan and manage an Azure AI solution 1.3 Manage, monitor, and secure AI systems

An Azure AI Search indexer shows the status Success after its nightly run, but agents…

Study guide skill: Monitor data ingestion quality, search index health, and relevance performance

Question B3-004

An Azure AI Search indexer shows the status Success after its nightly run, but agents can't find several documents that exist in the data source. What should you check first?

Select one answer.

  • A Nothing in the indexer: a Success status means every document was indexed, so the problem must be in the agent
  • B Reset the indexer, because a reset reports why individual documents failed
  • C The Document processed count metric in Metrics Explorer, which lists the error message for each failed document
  • D The errors and warnings for that run in the indexer's execution details (or the Get Indexer Status response), because a run can succeed while individual documents fail ✓ Correct answer

Answer: D

An indexer run reports Success even when some documents have errors, as long as the number of errors stays below the indexer's Max failed items setting. Document-specific problems are listed in the Errors and Warnings fields of each execution, which you can see in the portal or retrieve with Get Indexer Status. Azure Monitor metrics show counts of processed documents and skill invocations, not per-document errors, and a reset only clears change tracking so the next run reprocesses everything.

Deep dive

How it works

An Azure AI Search indexer reports status per run, and Success doesn't mean every document made it. A run is successful even when individual documents have errors, as long as the number of errors stays under the indexer's Max failed items setting. Only when errors exceed that limit does the run show Failed and indexing stop.

So when documents are missing after a "successful" run, start with the errors for that run:

  1. In the portal: select the indexer, then the run's status, to see start and end times plus detailed errors and warnings for that run.
  2. With REST: call Get Indexer Status. The response has the overall status, the last (or in-progress) run, and the history of recent runs, each with its errors and warnings.

The error tells you what to fix: a document too large, a failing skill, a field mapping problem, or a parsing issue. Fix the cause, then let the next run pick the documents up.

The other options, and when you'd use them

  • Success means everything was indexed: No. A run can succeed while individual documents have errors, if they're under Max failed items.
  • Reset the indexer: A reset clears the indexer's change tracking state so the next run processes everything in full. It doesn't report why documents failed.
  • Metrics Explorer: Azure Monitor metrics show the number of indexer-processed documents and skill invocations. That's for progress and alerts, not per-document error messages.

Exam tip

Indexer says Success but documents are missing = read that run's errors and warnings (portal execution details or Get Indexer Status). Remember Max failed items.

References

Last reviewed:

Ready to test yourself? The quiz asks this and every other AI-103 question, with instant feedback. Practice this in a quiz →

All Plan and manage questions · AI-103 study guide

Thanks for your feedback!