QuestionQ8

Implement machine learning model lifecycle and operations

A customer-facing web application uses a foundational model deployed through Microsoft Foundry.

A new model version must be introduced and validated without disrupting production traffic.

You need to deploy the new version by using a safe promotion strategy.

Which three actions should you perform, in sequence?

Drag & Drop
Deploy a new model version.
Create an AI gateway route version.
Shift 100% of traffic to the new version.
Validate the behavior of the new version.
Update the model settings on the production route.
Explanation

A safe promotion (blue-green / canary) strategy follows a deploy-validate-then-cut-over sequence. First you deploy the new model version alongside production while it receives no live user traffic, so production is undisturbed. Next you validate its behavior using Foundry evaluations, shadow/canary traffic, and quality gates before it serves real users. Only after it passes validation do you promote it by shifting 100% of traffic to the new version, completing the cutover with rollback still available. 'Create an AI gateway route version' and 'Update the model settings on the production route' are distractors that either skip the actual promotion step or would change production directly.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!