Question

A cloud operations engineer configures a monitoring dashboard for a customer-facing API. The dashboard displays CPU utilization aggregated as a 5-minute average and shows steady values around 45%. However, users report intermittent timeouts, and support tickets indicate slow responses during brief periods. The underlying compute instances autoscale on CPU, but scaling rarely triggers. What is the MOST likely reason the dashboard fails to reveal the problem, and what should the engineer change?

Accepted Answer

The average aggregation smooths out short-lived spikes; the engineer should add a maximum (or high-percentile) aggregation over a shorter period to expose transient saturation.

Answer

The metric collection interval is too short; the engineer should increase it to 15 minutes so trends are clearer.

Answer

CPU is the wrong metric entirely; the engineer should replace it with total request count only.

Answer

The dashboard time zone is misconfigured; the engineer should align it to UTC to match the incident reports.