Question 310
A monthly reporting job scores 30 million records overnight without per-row API latency requirements. What serving option is appropriate?
Explanation
Correct Answer
Batch endpoints are designed for asynchronous bulk scoring rather than a continually invoked low-latency API.
Incorrect Answers
Use online inference for synchronous responses or treat batch endpoint calls as asynchronous jobs. It addresses another case: A batch endpoint processes files from a protected data store but users expect an immediate JSON prediction response for each file. For that case, the distinguishing justification or acceptance check is: Batch endpoints execute workloads asynchronously and return job status, not interactive predictions in the immediate request response. The present case instead states: A monthly reporting job scores 30 million records overnight without per-row API latency requirements. It calls for Use an Azure Machine Learning batch endpoint or batch inference job. Showing success for the other case would not establish: Batch endpoints are designed for asynchronous bulk scoring rather than a continually invoked low-latency API.
The alternative, Configure supported private network access and authentication for the managed serving endpoint, belongs to a different technical condition: A production system needs to keep inference traffic on approved private networks and avoid public access. There the source-specific mechanism or proof is: Serving interface exposure is independent from model quality; network controls must enforce the application's required perimeter. The condition asked about here is: A monthly reporting job scores 30 million records overnight without per-row API latency requirements. A suitable correction is Use an Azure Machine Learning batch endpoint or batch inference job; the targeted justification or result is: Batch endpoints are designed for asynchronous bulk scoring rather than a continually invoked low-latency API.
Move credentials to a protected secret service or use identity-based access where supported. That operation would be relevant if: A public application uses a single fixed model-serving key stored in source code. Its corresponding reason or verification is: Secrets in application source are easily copied and difficult to rotate independently of software releases. However the problem in this question is: A monthly reporting job scores 30 million records overnight without per-row API latency requirements. The requirement instead calls for Use an Azure Machine Learning batch endpoint or batch inference job; the question-specific check is: Batch endpoints are designed for asynchronous bulk scoring rather than a continually invoked low-latency API.