AZ-400 Practice Question: Design and implement build and release pipelines
You are managing a pipeline that deploys a microservices application to multiple Azure Kubernetes Service (AKS) clusters in different regions. You want to implement a progressive exposure strategy where the deployment first goes to a small cluster (canary), then to a medium cluster, and finally to all clusters. The deployment should be automated but with the ability to halt if errors occur. What should you use?
⚠ Common exam trap
Candidates often confuse manual approval gates (Option A) with automated deployment gates (Option B), assuming any 'gate' requires human approval, when in fact deployment gates can be fully automated based on health metrics.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Use deployment gates with evaluation of health metrics (e.g., error rate) before proceeding to the next stage.
Deployment gates in Azure Pipelines allow you to automatically evaluate health metrics (such as error rate, CPU usage, or custom metrics from Application Insights) before promoting a release to the next stage. This enables a progressive exposure strategy (canary → medium → all clusters) with automated rollback or halt if the metrics breach thresholds, without requiring manual intervention.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Use manual approval gates between stages.
Why it's wrong here
Manual approval gates require a human to actively review and approve the release at each stage, which interrupts the automated pipeline flow and leaves progression dependent on human response rather than on automated health signals. This is contrary to the goal of automated progressive exposure and does not provide continuous validation of application metrics.
- ✓
Use deployment gates with evaluation of health metrics (e.g., error rate) before proceeding to the next stage.
Why this is correct
Deployment gates automatically evaluate predefined health metrics, such as error rate, latency, or availability, before allowing the pipeline to proceed to the next stage. They continuously assess these signals and can halt or fail the deployment if thresholds are exceeded, enabling safe, automated progressive exposure across clusters without manual intervention.
- ✗
Configure a rolling deployment strategy on each cluster.
Why it's wrong here
Rolling deployment strategy controls how new versions of pods are rolled out within a single Kubernetes cluster (e.g., maxSurge/maxUnavailable), ensuring availability during the update. It does not address the multi-cluster progressive exposure scenario, where the concern is validating health metrics before routing traffic to a new cluster. While a rolling update can minimize downtime within a cluster, it lacks the automated health-gate evaluation needed to decide whether to promote the release to the next cluster stage.
- ✗
Use a manual validation step in the pipeline.
Why it's wrong here
A manual validation step requires a team member to manually verify the deployment's behavior and then trigger the next stage, which is time-consuming and does not provide real-time, automated assessment of health metrics. It lacks the continuous monitoring and automated halt capability of proper deployment gates, making it unsuitable for safe progressive exposure.
Visual reference
Go deeper
Related to this question
Learn chapter
Introduction to DevOps and Azure DevOps
Key term
Application Insights
Application Insights is an Azure monitoring service that helps developers detect, diagnose, and understand issues in live web applications by collecting telemetry data like requests, exceptions, and performance counters.
Key term
Anthos
Anthos is a Google Cloud platform that lets you run applications consistently across different computing environments, like on-premises data centers and multiple public clouds.
About these practice questions
Courseiva writes every AZ-400 question from scratch — 823 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AZ-400 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AZ-400 exam.