Free ISTQB Certified Tester Advanced Level Technical Test Analyst Sample Questions

Correct answer: B — A requirement that has not been fully tested.

• A. Incorrect. A specific test case designed to address a risk is a test condition or test design artifact, not the risk itself. Test cases are created to mitigate or reveal risks, but they are not the definition of a test risk.
• B. Correct. A test risk is a product risk associated with a quality characteristic that may not be adequately delivered by the system. In ISTQB terms, this is often framed as the possibility that a requirement or aspect of the product has not been sufficiently tested, leading to potential failure or loss.
• C. Incorrect. An attribute of a test object that poses a risk is closer to a risk factor or source of risk, but it is not the definition of a test risk. The risk is the potential negative impact, not the attribute itself.
• D. Incorrect. A defect found during test execution is the result of testing, not a risk. A defect is an actual problem in the software, whereas a test risk is a potential problem or area of concern before or during testing.

Correct answers: B, D, F — Improved focus on critical areas of the software; Better alignment of testing with business priorities; More efficient use of testing resources

• A. Incorrect. Risk-based testing does not guarantee elimination of low-risk tests; it prioritizes effort based on risk. Some low-risk tests may still be executed for coverage, regression, or compliance reasons.
• B. Correct. A key benefit of risk-based testing is focusing testing effort on the most important and risky areas of the software, improving defect detection where impact is greatest.
• C. Incorrect. No testing approach can guarantee detection of all defects. Risk-based testing improves detection in critical areas but cannot prove the absence of defects.
• D. Correct. Risk-based testing aligns testing priorities with business impact and risk exposure, ensuring valuable functionality receives appropriate attention and making testing more relevant to stakeholders.
• E. Incorrect. Risk-based testing can reduce risk but cannot eliminate all project risks. Many risks (schedule, staffing, requirements) remain outside the scope of testing.
• F. Correct. By concentrating resources where risk is highest, risk-based testing helps use testing time, effort, and tools more efficiently, avoiding disproportionate effort on low-impact areas.

Correct answer: A — It may miss risks not on the checklist, especially new or technology-specific ones.

• A. Correct. Risk checklists are based on predefined lists of risks, which may overlook risks not already included, especially those related to new technologies or specific implementations. This is a known limitation of checklist-based approaches, as they cannot cover all possible risks.
• B. Incorrect. Checklists are typically used to streamline risk identification and reduce effort, not to require extensive brainstorming. While stakeholder input can be useful, it is not an inherent requirement of checklist-based risk identification.
• C. Incorrect. Checklists can be applied to safety-critical systems as part of a structured risk identification approach. The limitation is incompleteness, not unsuitability for such systems.
• D. Incorrect. Checklists can cover both functional and non-functional risks if designed appropriately. Their main weakness is being narrow or outdated, not restriction to functional risks only.

Correct answer: B — Run the simpler load test first, and only if problems are found, then consider the full test.

• A. Incorrect. Although the risk level is high, risk-based testing emphasizes selecting tests that provide the best information for the least cost. Jumping directly to the full performance test may be costly without initial evidence that a simpler test is insufficient. An incremental approach is often more appropriate.
• B. Correct. A simpler load test can act as an efficient first step to gather useful evidence about performance risk before committing to a more expensive full performance test. This reflects risk-based testing by balancing risk, cost, and available information. If problems are found, further testing can be justified.
• C. Incorrect. Unit tests are not suitable for assessing end-to-end latency or realistic system performance under load. They may help with code correctness, but they do not address the identified performance risk in the order routing module, especially under high demand.
• D. Incorrect. A code review can help find certain defects, but it is not a substitute for performance testing when the risk concerns runtime behavior such as latency. The risk here is about how the module behaves under load, which requires dynamic performance evaluation.

Correct answer: B — Two test cases are needed to cover all executable statements.

• A. Incorrect. A single test case cannot cover both branches of the if-else statement; one branch will be missed. Also, dead code is not executable and does not require coverage.
• B. Correct. Two test cases are sufficient: one with x > 0 (e.g., x = 1) to execute the 'Positive' branch, and one with x ≤ 0 (e.g., x = 0) to execute the 'Non-positive' branch. The dead code under 'if (false)' is never executed and does not affect statement coverage.
• C. Incorrect. Three test cases are unnecessary because there are only two reachable executable statements (the two print statements in the if-else). The dead code cannot be covered by any test case.
• D. Incorrect. Statement coverage focuses on executable statements only; dead code is ignored. Therefore, full statement coverage is achievable with a finite number of test cases.

Correct answer: B — 100%

• A. Incorrect. 66% would imply only two out of three statements were covered, but the test suite executes all three statements (S1, S2, S3) at least once, resulting in 100% coverage.
• B. Correct. Statement coverage is the percentage of executable statements executed at least once. Tests A, B, and C collectively execute S1, S2, and S3, so 3 out of 3 statements are covered, yielding 100% coverage.
• C. Incorrect. 33% would imply only one statement was executed, but all three statements are executed by the test suite.
• D. Incorrect. 50% would imply 1.5 statements were executed, which is not possible. All three statements are executed, so coverage is 100%.

Correct answer: B — To ensure that every executable statement runs at least once.

• A. Incorrect. This describes decision/branch coverage, which checks that each decision branch executes both true and false outcomes. Statement coverage does not focus on decision branches.
• B. Correct. Statement coverage aims to ensure that every executable statement in the code is executed at least once during testing. This is the basic objective of statement testing in white-box test techniques.
• C. Incorrect. This describes condition coverage or modified condition/decision coverage (MC/DC), which focuses on testing individual conditions in compound decisions. It goes beyond simple statement coverage.
• D. Incorrect. This describes path coverage, which requires exercising all possible paths through the control flow graph. Path coverage is generally more extensive and often not feasible for non-trivial code. It is not the same as statement coverage.

Correct answer: A — 100% decision coverage guarantees 100% statement coverage.

• A. Correct. In ISTQB terminology, decision coverage subsumes statement coverage. Achieving 100% decision coverage (testing all decision outcomes) requires that all statements in the code are executed, thus guaranteeing 100% statement coverage. However, the converse is not true: 100% statement coverage does not guarantee 100% decision coverage.
• B. Incorrect. 100% statement coverage ensures every statement is executed at least once, but it does not require testing both outcomes of decisions (e.g., both true and false branches). Therefore, it does not guarantee 100% decision coverage.
• C. Incorrect. Statement and decision coverage are related; decision coverage is a stronger criterion that subsumes statement coverage. They are not independent metrics.
• D. Incorrect. Decision coverage is actually a stronger (not weaker) criterion than statement coverage because it requires testing all decision outcomes, which inherently includes covering all statements.

Correct answer: B — Branch coverage

• A. Incorrect. Statement coverage ensures each executable statement is executed at least once, but does not guarantee that decisions are taken in both outcomes or that conditions are properly exercised. It is weaker than branch coverage for decision logic.
• B. Correct. Branch coverage exercises each decision outcome, providing better control-flow testing than statement coverage. It is commonly selected as a practical balance between testing effectiveness and effort.
• C. Incorrect. MC/DC is a much stricter criterion, typically used in high-integrity or safety-critical contexts. It requires demonstrating that each condition independently affects the decision outcome, which is more demanding than branch testing.
• D. Incorrect. Path coverage is generally impractical for non-trivial code due to the exponential number of paths from loops and branching. It is not the preferred technique for routine white-box test selection.

Correct answer: B — Branch coverage, to test decision logic within each service.

• A. Incorrect. Statement coverage only checks that each statement is executed at least once, but it does not guarantee thorough testing of decision logic within a microservice. The small size of microservices does not make statement coverage more suitable; it is still too basic for verifying internal control flow.
• B. Correct. Branch coverage tests each decision outcome (true/false), which is essential for verifying decision logic within a service. It is more effective than statement coverage for white-box testing of control flow and logical behavior.
• C. Incorrect. Service integration coverage is not a standard white-box test technique according to ISTQB. Testing communication protocols falls under integration testing, not white-box structural coverage.
• D. Incorrect. Path coverage is often impractical due to path explosion, especially across inter-service calls. It is not a realistic choice for typical internal logic testing in microservices.

Correct answer: B — Inject a JavaScript snippet into the comment field, then use the API's GET endpoint to retrieve the comment and check if the script is returned unsanitized in the JSON.

• A. Incorrect. While this approach checks for reflection of the script in the response, the presence of the script in the API response does not necessarily confirm an XSS vulnerability because the data may be safely encoded or consumed in a non-executable context (e.g., JSON). XSS risk depends on whether the payload executes in a browser, not just on reflection in the response body.
• B. Correct. This method directly tests the API's handling of unsanitized input by injecting a script and verifying that the stored comment is returned without encoding in the JSON response. If the API returns the script unsanitized, it indicates a stored XSS vulnerability, as a client consuming this data could later render it in a browser without safe encoding.
• C. Incorrect. This approach tests the web interface rather than the API itself. While it can confirm XSS execution, it does not isolate the API's role in preserving unsafe input. The question asks for API testing, so the most direct method is to examine the API's raw response for unsanitized data.
• D. Incorrect. Fuzz testing with random strings is useful for robustness but not effective for detecting XSS, which requires specific malicious payloads. XSS vulnerabilities typically do not cause server errors; they are client-side issues. Monitoring error rates is unreliable for identifying XSS flaws.

Correct answer: A — Test each service in isolation with mocks, then perform a single end-to-end test to verify the complete message flow.

• A. Correct. This strategy combines isolated service tests using mocks (which help localize failures) with a single end-to-end test that validates the real integration across the complete message flow. It is efficient and balances granular verification with end-to-end confidence.
• B. Incorrect. While this tests the full flow in a realistic environment, it is slow, brittle due to timing dependencies, and does not provide the isolation benefits that make debugging easier. It can be used as a supplementary test but is not the most appropriate primary strategy.
• C. Incorrect. These are unit-level tests that verify individual service behavior in isolation. They do not validate the interactions between services or the actual message flow across the queue, which is the key risk in a distributed workflow.
• D. Incorrect. Simulating the queue allows focused testing of message handling logic, but it does not verify the real messaging infrastructure or the complete end-to-end flow across all services. It is useful for unit testing but insufficient for verifying end-to-end message delivery.

Correct answer: C — The result (T/F) of evaluating a decision expression.

• A. Incorrect. This describes the set of all possible values of a condition, which is the domain of a condition, not the decision outcome. A decision outcome is the result of evaluating the entire decision expression, not individual condition values.
• B. Incorrect. The path taken is a consequence of the decision outcome, not the outcome itself. Decision outcome refers to the result of evaluating the decision expression, which determines the path.
• C. Correct. In decision testing, a decision outcome is the result (True or False) of evaluating the entire decision expression. It determines which path is taken and is the focus of decision testing to exercise both outcomes.
• D. Incorrect. The number of conditions is a structural characteristic of the decision, unrelated to the outcome. Decision outcome concerns the final evaluated result (True or False), not the count of conditions.

Correct answer: B — (Number of decision outcomes exercised / Total number of decision outcomes) × 100%

• A. Incorrect. This formula counts decisions executed, but decision coverage is based on the outcomes of decisions (e.g., true/false), not just reaching the decision point.
• B. Correct. Decision coverage measures the percentage of all possible decision outcomes that have been exercised. The formula is (Number of decision outcomes exercised / Total number of decision outcomes) × 100%.
• C. Incorrect. This is the formula for statement coverage, which measures the proportion of executable statements executed, not decision outcomes.
• D. Incorrect. This is the formula for path coverage, which measures the proportion of all possible paths through the code that have been tested, not decision outcomes.

Correct answer: A — Fewer distinct, observable test cases may be needed, since some combinations become indistinguishable at run time

• A. When an operand is short-circuited, its value cannot be independently observed for that branch, so combinations that only differ in the unevaluated operand collapse into one observable case, reducing the achievable count below the theoretical 2^N.
• B. Short-circuit operators do not make the decision untestable; testers simply account for which combinations are actually observable at run time.
• C. Short-circuiting reduces, rather than doubles, the number of distinguishable test cases because some combinations become unreachable or indistinguishable.
• D. Short-circuit evaluation does have an effect: it can make some of the theoretical 2^N combinations unreachable or indistinguishable, so fewer than 2^N distinct cases may be observable.

Correct answer: A — MC/DC, needing only nine test cases while still proving each condition independently affects the outcome

• A. MC/DC is the industry-preferred alternative in safety-critical domains such as avionics because it needs roughly N+1 test cases (nine here) while still demonstrating that each condition independently changes the decision outcome, giving strong confidence at a fraction of the cost of full combination testing.
• B. Statement coverage is a much weaker criterion that says nothing about condition combinations and would not provide the confidence in decision logic that the team needs.
• C. Basic condition coverage only requires each condition to independently reach true and false somewhere in the suite, without linking those outcomes to the decision result, so it gives weaker assurance than what safety-critical logic typically requires.
• D. Random testing has no structural guarantee of exercising specific condition combinations or proving independence of effect, making it unsuitable for the rigor DO-178C safety-critical decisions demand.

Correct answer: B — The count of linearly independent paths through the code.

• A. Incorrect. The total number of executable statements measures code size, not logical complexity. Cyclomatic complexity is based on control flow, not statement count.
• B. Correct. Cyclomatic complexity is defined as the count of linearly independent paths through the code, calculated from the control flow graph. Higher values indicate more decision points and increased testing effort.
• C. Incorrect. Duration of execution is a performance metric obtained through dynamic analysis, not a static complexity measure like cyclomatic complexity.
• D. Incorrect. Heap memory allocation is a resource usage metric, relevant for performance testing but not a measure of structural complexity or path count.

Correct answer: B — Permit the build to pass and alert developers while monitoring the trend.

• A. Incorrect. Removing the static analysis stage entirely would eliminate an important quality gate and reduce defect detection capability. A minor issue does not justify abandoning a practice that helps prevent future problems.
• B. Correct. If the issue is low-risk and does not block delivery, it is reasonable to let the build continue while informing developers so they can address it and track whether the trend worsens. This balances delivery speed with visibility and continuous improvement.
• C. Incorrect. Restricting analysis to the stable production branch is too late in the lifecycle and reduces the value of early defect detection. Static analysis is most effective when applied early and consistently across development changes.
• D. Incorrect. Replacing automated static analysis with manual inspection would make the process less scalable and less consistent. Manual review can complement automation, but it should not substitute for it as the primary detection mechanism.

Correct answer: B — System mode profile

• A. Incorrect. A user profile categorizes system usage based on user types or user behavior patterns, not operational states like startup or idle. It focuses on actions performed by different user groups rather than the system's internal operating states.
• B. Correct. A system mode profile categorizes system behavior into distinct operational states or modes such as startup, idle, busy, or shutdown. It models how the system transitions between these states, describing behavior in different modes of operation.
• C. Incorrect. A functional profile describes the distribution or usage of functions or features, often by importance or frequency, rather than separating behavior into operational states. It is about what the system does, not the mode it is in.
• D. Incorrect. A reliability growth profile models how reliability improves over time as faults are detected and removed during testing or operation. It does not categorize the system into states like startup, idle, and busy.

Correct answer: A — Operational profiles provide expected usage input to drive reliability growth models

• A. Correct. Operational profiles define the expected usage scenarios, frequencies, and probabilities of system inputs, which are essential for driving reliability growth models. This input helps prioritize the most realistic and impactful usage patterns when modeling and measuring failure behavior over time.
• B. Incorrect. Reliability growth testing does not make operational profiles obsolete; in fact, it depends on them. Operational profiles help ensure the test effort reflects real-world use so that reliability improvements are measured against realistic conditions.
• C. Incorrect. Operational profiles are not limited to performance testing; they are also relevant for reliability testing, including reliability growth testing, where they model how the system will be used in the field.
• D. Incorrect. Reliability growth models do not automatically generate operational profiles. Operational profiles must be developed from usage analysis, field data, expert judgment, or customer information before they can be used in reliability growth testing.

Correct answer: B — Inter-service coupling complexity.

• A. Incorrect. The number of HTTP requests per service is more related to performance or efficiency testing rather than maintainability. It measures communication volume but does not directly assess ease of modification.
• B. Correct. Inter-service coupling complexity is a key maintainability metric because high coupling makes services harder to change, test, and evolve independently. Lower coupling improves understandability, modifiability, and long-term maintainability.
• C. Incorrect. Average request latency is primarily a performance metric measuring response speed, not maintainability. It does not indicate how easily the system can be modified or maintained.
• D. Incorrect. Deployment frequency is mainly associated with release agility and DevOps efficiency. While it can indirectly reflect maintainability, it is not a direct measure of maintainability itself.

Correct answer: B — Depth of inheritance.

• A. Cyclomatic complexity is a code complexity metric that measures the number of linearly independent paths. It is primarily used for testability and complexity analysis, but it can indirectly indicate maintainability issues. However, it is not a direct maintainability metric according to ISTQB.
• B. Depth of inheritance is a classic maintainability metric. It measures the depth of a class hierarchy; deeper hierarchies can increase complexity and reduce maintainability, making it a key metric for assessing design maintainability.
• C. Comment quality index is not a standard maintainability metric in ISTQB. While comments can aid readability, there is no widely recognized metric for comment quality; Instead, comment density or ratio may be used.
• D. Lines of code per module is a size metric that can provide an indirect indication of maintainability, as larger modules are generally harder to understand and modify. However, it is not specifically a maintainability metric; it is more a measure of module size.

Correct answer: B — Sensitive Data Exposure

• A. Incorrect. Missing authentication is a security weakness that allows unauthorized access, but it does not specifically address exposure of protected data once access is gained. It concerns entry point control rather than data confidentiality.
• B. Correct. Sensitive Data Exposure occurs when confidential information is not properly protected and can be revealed to unauthorized parties. This directly impacts the confidentiality quality characteristic.
• C. Incorrect. Insecure Direct Object References (IDOR) is an access control flaw where users can access objects they should not, but it is not primarily about exposure of sensitive data without adequate protection; it is about directory/parameter manipulation.
• D. Incorrect. Cross-Site Request Forgery (CSRF) tricks a user's browser into executing unwanted authenticated requests. It relates to session and request integrity, not primarily to unauthorized disclosure of sensitive data.

Correct answer: B — Implementing multi-factor authentication

• A. Using TLS for all communication protects data in transit from interception and tampering, but it does not directly verify a user's identity or prevent unauthorized access to accounts. It is an important security measure, but not the most effective control for this specific purpose.
• B. Implementing multi-factor authentication adds an additional verification factor beyond just a password, making it much harder for attackers to gain unauthorized access even if credentials are compromised. It is one of the most effective controls for protecting user accounts.
• C. Allowing weak passwords increases the risk of brute-force attacks and credential guessing, which weakens account security. This is the opposite of a control that prevents unauthorized access.
• D. Enforcing account lockout after failed attempts helps mitigate brute-force attacks by limiting repeated login attempts, so it is a useful protective control. However, it is generally less effective than multi-factor authentication for preventing unauthorized access to user accounts.

Correct answer: B — Validate shopping cart data integrity on server side

• A. Incorrect. Multi-factor authentication authenticates users but does not directly prevent manipulation of cart data. It addresses identity theft rather than data integrity.
• B. Correct. Server-side validation verifies the integrity and authenticity of cart data before processing, directly preventing malicious modifications from being accepted.
• C. Incorrect. HTTPS encrypts data in transit, protecting against eavesdropping, but it does not verify the validity of the data after it reaches the server. Tampered data can still be sent over HTTPS.
• D. Incorrect. Vulnerability scans identify security weaknesses but do not provide real-time protection against cart tampering. They are a preventive measure, not a direct control.

Correct answer: A — MTBF applies to repairable systems; MTTF to non-repairable systems.

• A. Correct. MTBF is specifically used for repairable systems, as it measures the average operating time between failures. MTTF is used for non-repairable systems, measuring the average time until the first (and only) failure occurs. This distinction is fundamental in reliability engineering.
• B. Incorrect. MTBF does not include repair time; it refers only to the average operating time between failures. Repair time is considered separately in metrics such as Mean Time To Repair (MTTR).
• C. Incorrect. Both MTBF and MTTF can be applied to hardware and software systems. The distinction is based on whether the system is repairable (MTBF) or non-repairable (MTTF), not on the type of system.
• D. Incorrect. Both MTBF and MTTF are measures of reliability, not availability. Availability is typically measured using metrics like uptime percentage or MTTR, though MTBF and MTTF contribute to availability calculations.

Correct answer: C — Robustness testing

• A. Incorrect. Functional correctness testing verifies that the software meets its specified functional requirements. While correct functionality is important, it does not directly assess the system's ability to operate reliably over time or under stress, which is the focus of reliability testing.
• B. Incorrect. Performance testing evaluates speed, scalability, and resource usage, focusing on time behavior and efficiency. Although some reliability aspects like endurance may be observed during performance tests, performance testing is not primarily aimed at measuring reliability characteristics such as fault tolerance or recoverability.
• C. Correct. Robustness testing assesses the system's ability to handle invalid inputs, unexpected conditions, and stressful environments without failure. This directly addresses the reliability subcharacteristics of fault tolerance and recoverability, making robustness testing the most closely related type.
• D. Incorrect. Regression testing ensures that recent code changes do not introduce new defects in existing functionality. It is a maintenance activity and does not focus on evaluating the system's long-term stability, failure behavior, or ability to handle stress.

Correct answer: B — Check that date and number formats are parsed and displayed correctly for each locale.

• A. Verifying that translated text fits within UI elements is important for usability and layout, but it is not the most critical test for ensuring core functionality is not broken. This focuses on presentation rather than functional behavior.
• B. Date and number formats vary significantly across locales (e.g., MM/DD/YYYY vs. DD/MM/YYYY, comma vs. period as decimal separator). Incorrect parsing or display can break functionality, such as form submissions, calculations, or data storage. Testing these ensures the application handles locale-specific data correctly.
• C. Externalizing strings is a key internationalization practice and a prerequisite for localization, but it is a code quality check performed before testing. During portability testing, the focus is on runtime behavior; confirming externalization does not directly test whether functionality is broken after localization.
• D. Translation memory tools improve translation efficiency, but their alignment accuracy is a tool-specific concern that does not directly verify the application's functional correctness after localization. This is more of a process-efficiency check.

Correct answer: B — Convert the Oracle-specific data types and stored procedures to their MySQL equivalents.

• A. Incorrect. While verifying tool compatibility is relevant, it is not the first priority because the core technical challenge is adapting the database structures and logic (data types, stored procedures) to the target DBMS.
• B. Correct. Oracle-specific data types and stored procedures are often incompatible with MySQL, so they must be identified and converted first before a meaningful restore and test can occur. This addresses the fundamental portability issue.
• C. Incorrect. Creating a mirrored schema in MySQL is a necessary step but should follow the conversion of Oracle-specific elements. Without first addressing data type and logic incompatibilities, a mirrored schema alone does not ensure portability.
• D. Incorrect. Estimating network transfer time is a logistical concern, not a technical portability issue. The test's objective is to validate whether the backup data and database objects can be successfully moved and used on MySQL.

Correct answers: A, B, E — Compatible with existing test management.; Parallel execution across multiple environments.; Support for protocols and interfaces of SUT.

• A. Compatibility with existing systems ensures seamless integration and reduces implementation overhead, making it a key practical consideration for tool selection.
• B. Parallel execution capabilities improve efficiency and scalability, which are critical for technical testing in complex or large-scale systems.
• C. The tool's logo and color scheme are cosmetic and do not affect its technical suitability.
• D. Aesthetic appeal of the user interface is not a primary factor; technical capability is more important.
• E. Support for specific protocols and interfaces of the system under test is essential for the tool to interact correctly and perform effective testing.
• F. Popularity in non-technical domains does not guarantee suitability for technical testing; relevant criteria are technical capabilities and alignment with the system architecture.

Correct answer: B — Advise against function points; recommend expert judgment based on technical complexity.

• A. Function point analysis is a sizing technique for functional requirements and development effort, but it does not directly account for technical complexity, non-functional aspects, or risks that drive technical test effort.
• B. Expert judgment based on technical complexity is well suited for estimating technical test effort because it directly considers architecture, interfaces, data volumes, non-functional requirements, and technical risks—factors that are more influential than functional size.
• C. Story points are an agile relative estimation tool used primarily for development work and do not reflect the specific technical testing considerations such as test depth, tooling, or non-functional requirements.
• D. Code coverage metrics measure test thoroughness after execution, not effort drivers. Combining them with function points does not address the key factors of technical test effort like complexity and test techniques.

Correct answer: B — Analyze the detailed performance test results to isolate which component or transaction step contributed most to the increased response time.

• A. Incorrect. While an endurance test can identify memory leaks and gradual degradation, it is not the appropriate first step after a sudden response-time increase. Endurance tests are time-consuming and should be performed after isolating the root cause of the performance regression to avoid wasted effort.
• B. Correct. The first investigation should focus on analyzing detailed performance test results to isolate which component or transaction step contributed most to the increase. This directly pinpoints the bottleneck and supports efficient root-cause analysis, making it the most effective first step.
• C. Incorrect. Re-running the baseline test can help confirm the degradation, but it does not directly address the root cause. Reviewing deployment artifacts is a valuable secondary step after the problematic component has been identified, not the immediate first action.
• D. Incorrect. Conducting a load test with increased user load is useful for stress testing but does not address the immediate question of why response time increased after deployment. The first step should be to analyze and isolate the regression under comparable conditions, not to intensify the workload.

Correct answers: A, B — Heap memory usage trend; Garbage collection frequency and duration

• A. Correct. Heap memory usage is a primary indicator of a memory leak. A steady increase over time without returning to a stable baseline, even after garbage collection, directly confirms that memory is being allocated but not released.
• B. Correct. Garbage collection frequency and duration increase when the system struggles to reclaim memory due to a leak. Frequent and longer GC cycles indicate memory pressure, supporting the suspicion of a memory leak.
• C. Incorrect. The number of concurrent user sessions is a load characteristic, not a direct metric for memory leaks. While it may correlate with memory usage, it does not confirm whether memory is being leaked.
• D. Incorrect. Average response time may degrade due to memory pressure, but it is an indirect symptom. It cannot distinguish between memory leaks and other performance issues like CPU saturation or network latency.
• E. Incorrect. CPU utilization may increase due to heavy garbage collection, but it is not a primary or direct indicator of a memory leak. It can support diagnosis but does not show memory retention.
• F. Incorrect. Disk write throughput is unrelated to memory leaks, as memory leaks occur in RAM. This metric is relevant for I/O performance but not for confirming memory leaks.

Correct answer: B — Checks for design, coverage, and traceability aspects.

• A. Incorrect. Checklists in reviews are not limited to syntactic correctness and formatting; they guide reviewers to identify a broader set of defects relevant to the work product, including design, coverage, and traceability.
• B. Correct. Checklists in reviews commonly cover important quality aspects such as design, coverage, and traceability. They help ensure consistency with requirements and systematic defect detection across multiple quality attributes.
• C. Incorrect. While performance test environment configuration may be checked in some reviews, checklists are not restricted to that area; they address a wider range of review objectives and defect types.
• D. Incorrect. Coding standards for test scripts may be part of a checklist in some contexts, but checklists are not limited to that focus. They support systematic defect detection across many quality attributes, not just coding standards.

Correct answer: B — During the individual checking phase to guide reviewers' focus.

• A. Incorrect. Checklists are not primarily used during the planning phase to schedule and assign reviewers. Planning focuses on logistics, roles, and scope, not the detailed guidance provided by checklists.
• B. Correct. Checklists are used during the individual checking phase to guide reviewers' focus, helping them systematically cover typical defect types, quality characteristics, and common problem areas before the review meeting.
• C. Incorrect. The follow-up phase is used to check that identified defects have been corrected and that review objectives are met. While checklists may be referenced, their primary role is not to verify fixes.
• D. Incorrect. The review meeting focuses on discussion and consensus, not on using checklists for documentation. By the meeting, reviewers should have already used checklists during individual preparation.