AZ-305 Practice Question 20: Migration Assessment First Step
Microsoft Azure Solutions Architect Expert (AZ-305) · Test 2 · question 10 of 10
A publisher plans to exit its datacenter, which runs roughly 800 VMware virtual machines. The application owners have left the company and no one can say which servers belong to which application. Before any migration wave is scheduled, the CFO wants a sized, costed Azure target estate, and the migration lead wants to know which servers must move together so that a wave does not split an application in half. The VMware team will not allow agents to be installed on the guest operating systems.
What should you do first?
Show answer and explanation
Answer:
Discovery and assessment come before migration in the Cloud Adoption Framework, and they are what produce both artifacts this scenario demands. The Azure Migrate appliance discovers the VMware estate through vCenter Server and the assessment returns right-sized Azure targets with a monthly cost estimate. Agentless dependency analysis captures TCP connection data from the discovered servers (polled by the appliance every five minutes, with no agent installed on any guest), and Microsoft's stated advantage for it is precisely the migration lead's question: it lets you group interdependent servers into applications and identify servers that must be migrated together, so nothing is left behind and there are no surprise outages after migration.
Option B migrates before it knows what it is migrating. Replication telemetry reports copy progress, not application relationships, and starting 800 replications commits egress and storage spend before anyone has validated sizing, cost, or wave composition.
Option C produces a number the CFO can be shown and little else. A same-size match ignores actual utilization, so it usually over-provisions; and a CMDB no one has maintained since the application owners left cannot answer the dependency question, which is the reason the estate is being analyzed in the first place.
Option D would eventually produce a dependency map, but it breaks the explicit constraint: agent- based analysis requires an agent on each source server. It is also the slower path to the same picture and still produces no sizing or cost assessment, which agentless discovery in Azure Migrate delivers as part of the same exercise.
Reference: Dependency analysis in Azure Migrate