AZ-400 Question 210
Select 2You are a DevOps engineer for a company that has multiple Azure DevOps pipelines running concurrently. While your build and release processes work, you notice that pipeline concurrency spikes are causing extended queue times for critical applications and driving up costs on your paid parallel job capacity. You decide to optimize concurrency to balance performance and cost. Which two approaches should you take to achieve this goal?
- A
Combine multiple independent pipelines into a single pipeline without defining distinct stages or dependencies
- B
Use pipeline-level concurrency limits in your YAML configuration to restrict how many runs can execute at once
- C
Configure scheduling and triggers so that lower-priority pipelines run during off-peak hours or only after critical runs complete
- D
Increase your paid parallel job count in Azure DevOps to ensure no pipeline ever waits in the queue
Show answer and explanation
Correct answers: B, C
Explanation
To optimize pipeline concurrency for performance and cost, Azure DevOps allows limiting how many jobs run in parallel by configuring concurrency in YAML pipelines and by effectively scheduling or prioritizing less critical pipelines. This approach prevents resource contention while reducing the risk of incurring higher charges for additional parallel jobs. For more details, refer to the official Microsoft Azure DevOps documentation on pipeline concurrency and scheduling: https://learn.microsoft.com/azure/devops/pipelines/process/concurrency.
- A. Incorrect.
Incorrect. Simply merging pipelines into one without clear stage definitions can reduce visibility into failures and does not inherently control concurrency per pipeline stage. This approach may complicate troubleshooting and does not directly address cost optimization.
- B. Correct.
Correct. Azure DevOps allows you to specify concurrency limits in your YAML configuration. For example, using the 'limit' keyword with 'runsOn' or configuring separate stages with different concurrency settings can help reduce simultaneous runs and manage cost.
- C. Correct.
Correct. Scheduling non-critical pipelines to run during periods of lower resource usage or after critical pipelines finish ensures more efficient resource allocation. This approach helps you pay only for concurrency when it� truly needed, balancing performance and cost.
- D. Incorrect.
Incorrect. Increasing the parallel job count will likely resolve queue delays but at a higher ongoing cost. This does not represent a cost optimization strategy; it merely throws more resources at the problem without addressing concurrency management.