Microsoft DevOps Engineer Expert AZ-400 Practice Question

Your organization uses multi-stage YAML pipelines in Azure DevOps to deploy dozens of microservices to a production environment. Leadership wants a delivery metric that shows the number of successful deployments to production each day over the last 30 days. You must deliver this metric in Power BI with minimal custom coding, while allowing product owners to filter by service (pipeline name).

Which approach should you use to obtain the data that drives the metric?

Connect Power BI to the Azure DevOps Analytics OData feed and query the PipelineRuns entity set.
Enable Continuous Export in Application Insights for each service and run a Kusto query in Power BI against the exported customEvents.
Create a custom Azure DevOps Analytics view that exposes PipelineRun data, then connect Power BI through the built-in Data Connector.
Query Azure Boards work items using the OData WorkItems entity and count those that have the state 'Closed' and the tag 'Production Deployment'.

Report Issue

Answer Description

The correct approach is to connect Power BI directly to the Azure DevOps Analytics OData feed. The PipelineRuns entity set within the Analytics data model contains authoritative data for each execution of a YAML pipeline, including the outcome, environment details, completion date, and pipeline name. This allows for the creation of a report on successful production deployments per day with filtering capabilities, fulfilling the requirements with minimal custom work.

Creating an Analytics View is incorrect because this feature is limited to Azure Boards work item data and does not support pipeline data. Using Azure Boards work items with a specific tag is not an authoritative source for deployment outcomes and relies on manual process adherence. Enabling continuous export from Application Insights would require custom instrumentation in every pipeline to send deployment events, which violates the 'minimal custom coding' requirement.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.