AI-300 Practice Question 52: Private Workspace DNS
Microsoft Machine Learning Operations Engineer Associate (AI-300) · Test 6 · question 2 of 10
A company created an Azure Machine Learning workspace with a private endpoint in a hub virtual network. Data scientists connect from the corporate network over a VPN, but studio and SDK calls fail because the workspace name resolves to a public IP address. The private endpoint itself shows as approved and healthy. What should the network engineer do?
Show answer and explanation
Answer:
A private endpoint only helps if the workspace fully qualified names resolve to its private IP address. Azure Machine Learning uses private DNS zones such as privatelink.api.azureml.ms and privatelink.notebooks.azure.net. Those zones must be linked to the virtual network, and the corporate DNS servers must forward queries for the workspace names to a resolver that can read them (option A).
- B is wrong because an IP allow list applies only to the public endpoint. It does nothing for name resolution and would keep traffic on the public path instead of the private one. - C is wrong because private endpoints are network interfaces in an Azure virtual network. They cannot be created in an on-premises network, and a second one would not change how names resolve. - D is wrong because RBAC decides what an authenticated user may do. It has no effect on which IP address a DNS name resolves to.
Reference: Use a custom DNS server with an Azure Machine Learning workspace