What practice helps identify performance regressions during testing?

• A. Compare to baseline metrics under realistic load, and log deviations.
• B. Only test at startup.
• C. Ignore CPU usage.
• D. Rely only on bug reports.

Answer: (A)

Comparing current performance to a known good baseline under realistic load is the way to spot regressions. By running tests that mimic real-world usage and measuring metrics like response time, throughput, and error rates, you can see when changes cause performance to slip below the baseline. Logging deviations gives concrete data on when and how the system behaves differently, making it possible to identify and triage regressions quickly.

This approach catches issues that functional tests can miss — for example, a change that doesn’t break correctness but slows responses under typical user load, or causes higher resource use over time. It also provides objective thresholds to trigger alerts rather than relying on sporadic bug reports or isolated checks.

Other approaches fall short because testing only at startup might miss long-running or steady-state regressions, ignoring CPU usage overlooks bottlenecks, and relying solely on bug reports often ignores performance problems that users may not immediately report.