CT-PT ISTQB Certified Tester - Performance Testing Question and Answers

Ifmultiple virtual users execute the same script while logged in as the same user, it can causedata conflicts, session collisions, or cache interference, leading toinconsistent performance test results.

Option A (Concurrent scripts accessing the same data)may causedata locking issues, but this is atest setup concern rather than an inconsistency issue.

Option B (Sequential scripts accessing the same data)does not typicallycause inconsistenciessince they execute in order.

Option D (Verification with management)is unrelated totechnical inconsistencies.

Crowd-based load generationrelies on real users from different locations, makingtest conditions highly variable. This makes itdifficult to reproduce the same test scenario multiple timesunder identical conditions.

Option B (Less sensitive to changes)is incorrect becausecrowd-based load is highly sensitive to external conditions(network, device types, etc.).

Option C (Suitable for mainframe applications)is incorrect becausecrowd-based load testing is typically used for web and cloud applications, not mainframes.

Option D (More precise than other methods)is incorrect becausecrowd-based load testing lacks precision due to real-world inconsistencies.

Performance testing istime-sensitiveand must be designed tofit within the constraintsof the project timeline. If performance tests take too long to execute, they may not be feasible within a given sprint or development cycle. The results should be reproducible, meaning the same test on an unchanged system should yield the same results (making option C incorrect). While stakeholder expectations are important, performance tests should be objective and based on defined benchmarks rather than subjective expectations (making option D incorrect).

Before running the next set of performance tests, it is crucial tounderstand the differences (delta) between the previous test results and the expected improvements. This ensures that:

The changes made by developers are properly validated.

Performance objectives are aligned with measurable improvements.

Option B (Adjust stakeholder expectations)is incorrect, as performance testing should focus on real data rather than adjusting expectations.

Option C and D (Realign objectives for milestones or time-to-market goals)are incorrect becausetest objectives should focus on technical accuracy rather than schedule constraints.

One of the mostcritical test acceptance criteriain performance testing is toensure that the hardware in the test environment is comparable to production. Differences inCPU, memory, disk I/O, or network infrastructurecandistort performance results.

Option A (Convincing stakeholders to set goals)is a planning activity, not an acceptance criterion.

Option B (Describing the system under test)is important but does not directly affect test acceptance.

Option D (Comparing baselined metrics)is useful, butwithout a comparable test environment, baseline metrics may be misleading.

Test implementationis the stage wheresmaller test components are created and assembled into larger test scenarios. This phase includesscripting, test data preparation, and configuring load profiles.

Option A (Test completion)involves analyzing results, not building tests.

Option B (Test planning)focuses onstrategies and objectives, not test creation.

Option C (Test design)definestest scenarios and cases, but the actual building happens inimplementation.

One of the keyperformance risks in mainframe architecturesisslow hard disk speed, which canimpact transaction processing, batch jobs, and database access times.

Option A (Inconsistent loads via APIs)is more relevant todistributed or cloud architecturesrather than mainframes.

Option C (Misconfigured host machine)is aconfiguration issue, not a typical performance risk.

Option D (Lack of RAM)is less common in mainframes, which typically useoptimized memory management.

To ensure anaccurate operational profile, it is crucial tovalidate and adjust it with key stakeholders before executing tests. This step ensures that the workloadrealistically represents actual user behavior.

Option A (Executing the test first and adjusting later)is incorrect becauseadjustments should be made before executionto avoid misleading results.

Option C (Aggregating profiles into a single generic profile)removesimportant distinctions between different user types, reducing test accuracy.

Option D (Creating separate profiles and using equal proportions)does not match the actualpower purchaserbehavior, which requires arealistic representation of their higher-volume transactions.

Protocol-level scriptsmeasure theactual response times of network requests, making them ideal foranalyzing network latency, bandwidth issues, and server response times.

Option A (GUI scripts)measure user interactions, not network timing.

Option B (API scripts)measure API interactions but do not providedetailed network response time analysis.

Option C (HTTP scripts)measure web requests but lack low-level network insights.