AI-300 Practice Question 92: Compute Instance Stop Schedule

Microsoft Machine Learning Operations Engineer Associate (AI-300) · Test 10 · question 2 of 10

Design and implement an MLOps infrastructureeasy

Developers often leave compute instances running overnight, sometimes with a notebook kernel still open. The team wants every compute instance to stop at 8 PM on weekdays, even if a kernel is still running. What should they configure?

Select one answer
Show answer and explanation

Answer:

  • A. A schedule on each compute instance that stops it at 8 PM

A compute instance can have a schedule that starts or stops it at set times. A stop schedule ends the instance at the set time, whether or not a kernel is open (users get a notification beforehand). Idle shutdown is different: an instance is idle only when there is no active Jupyter kernel session and similar activity (option A).

- B is wrong because an instance with an active kernel session is not considered idle, so it would not be stopped at 8 PM. - C is wrong because a compute cluster is a different resource. It would not stop the existing compute instances. - D is wrong because removing permissions limits who can create compute, and does not stop instances that already exist.

Reference: Create an Azure Machine Learning compute instance

Practice Test 10 ›

Study guides for AI-300