‹ All AI-300 questions

AI-300 Practice Question 51: Workspace Level Quota

Microsoft Machine Learning Operations Engineer Associate (AI-300) · Test 6 · question 1 of 10

Design and implement an MLOps infrastructuremedium

A subscription has a development workspace and a production workspace in the same region. By default both draw on the same subscription quota for a GPU VM family. Large experiments in the development workspace sometimes use all of the quota, and scheduled jobs in the production workspace then fail to get nodes. What should the team do?

Select one answer
Show answer and explanation

Answer:

  • A. Set a workspace-level quota on the development workspace to cap its share of the VM family quota

Compute quota is set for the subscription per region and VM family, and by default all workspaces in the subscription share it. Workspace-level quotas let an administrator set a maximum for each workspace, so the development workspace can no longer use up the capacity that production needs (option A).

- B is wrong because roles control permissions and would not change how the shared quota is divided between the workspaces. - C is wrong because quota is applied per subscription and region, so a different resource group in the same subscription still draws on the same quota. - D is wrong because the production cluster can only obtain nodes that the shared quota still has available, so a higher maximum does not help.

Reference: Manage resources and quotas

Practice Test 6 ›

Study guides for AI-300