A worker fleet has low CPU but a growing work queue because each task spends time waiting on external I/O. Which autoscale signal best represents unmet demand?

Choose the single best answer.

A.Queue depth or work items per worker, with downstream-aware concurrency limits
B.CPU utilization alone with no backlog monitoring
C.The worker instance count alone, regardless of pending work
D.Completed-task count alone, excluding pending work and age

Show the answer and explanation

✓ Correct answer

Queue depth or work items per worker, with downstream-aware concurrency limits

Queue depth measures unmet demand when workers are I/O-bound. CPU can miss this bottleneck, instance count describes allocation rather than backlog, and completed-task count alone can hide accumulating work. Scaling must still protect downstream capacity. Source guide: PDF page(s) 45.

Read the source reference

This item is flagged for content review.

Read and review here, or open the app to practise and save your results on this device.

Practise in the app

Unofficial study aid; not affiliated with or endorsed by Microsoft. Original practice questions are not real examination questions and do not guarantee an exam result.