FreeExam

Test strategy

A QA engineer interview on choosing what to test and at which layer, designing test data and doubles, handling flaky tests, and judging release readiness.

These prompts are original practice. They are not questions from a named employer or a live hiring loop.

  1. 1.A bike-share app checks its fare rules, which depend on ride minutes and membership tier, only through browser tests that take 40 minutes. What is the best change?

  2. 2.What does the test pyramid recommend about the mix of automated tests?

  3. 3.A payroll service shipped a bug where an overtime query read the wrong column. Its unit tests mock the repository class and all passed. Which test would most likely have caught the bug?

  4. 4.A team has about 600 browser tests and 40 unit tests. Which problem is this shape most likely to cause?

  5. 5.A grain co-op prices loads with a pure function of moisture, grade, and weight. There are about 30 meaningful input combinations. Where should most of these checks live?

  6. 6.What does an end-to-end test give you that lower-layer tests do not?

  7. 7.A county library's catalog service calls a new hold-queue service. You need to know the catalog handles a 'queue full' response well. What is the cheapest reliable test?

  8. 8.You are testing a React sign-up form that shows validation messages. Which approach gives the most useful tests?

  9. 9.Which check belongs in a unit test rather than an integration test?

  10. 10.A clinic scheduler release is due in two weeks and you cannot test everything. What do you do first?

  11. 11.In risk-based testing, the level of risk for a feature is usually judged from which two things?

  12. 12.A tax-filing app's release changes how refunds are rounded and also changes a button color on the help page. You have one day. Where do you spend it?

  13. 13.A product manager asks what was not tested before a release. What is the best response?

  14. 14.Which fact about a ferry ticketing module should most raise its testing priority?

  15. 15.What belongs in the 'out of scope' section of a test plan?

  16. 16.A hotel booking site supports 12 payment providers, and 3 of them carry 95 percent of traffic. Test time is tight. What plan makes sense?

  17. 17.A gym sign-up form accepts ages from 18 to 65 inclusive. Which set of values gives the minimal boundary tests?

  18. 18.What does equivalence partitioning let a tester do?

  19. 19.A parcel service charges a flat rate under 1 kg, a per-kg rate from 1 to 5 kg, and freight rates above 5 kg. Weights are whole grams. Which test set is best?

  20. 20.A bakery's preorder form takes a whole-number quantity from 1 to 24. Which input belongs to a different invalid class than 30?

  21. 21.Why should a test usually include only one invalid input at a time?

  22. 22.A store gives a discount when the cart total is 'over $50'. The developer wrote total >= 50. Which test catches this?

  23. 23.A username must be 3 to 20 characters. Which set of tests is most complete?

  24. 24.A magazine renews subscriptions on the same day each month. Which dates are most worth including in tests?

  25. 25.Many tests share one seeded admin user. One test changes that user's password, and later tests start failing. What is the best fix?

  26. 26.A team wants to copy production data into the test environment. What is the main risk?

  27. 27.Hundreds of tests build a Patient object with 25 fields by hand. A new required field breaks 300 tests. What is the best fix?

  28. 28.A test loads a 5 MB fixture with 10,000 orders to check one tax rule. What is the main problem?

  29. 29.A test that uses randomly generated inputs fails once, and nobody can reproduce it. What should the team change?

  30. 30.A pension fund app needs realistic test data, but no real member records may leave production. What is the best approach?

  31. 31.A tool rental shop's late-fee test passes most days but fails on the last day of each month. What is the best fix?

  32. 32.What is the difference between a fake and a mock in testing?

  33. 33.A team refactors how a service calls its object-relational mapper (ORM). Behavior is unchanged, yet 80 heavily mocked tests break. What does this show?

  34. 34.A weather-alert service sends texts through a third-party SMS provider. How should automated tests handle that provider?

  35. 35.What does a consumer-driven contract test verify?

  36. 36.A mobile app's tests use a stub of the inventory API that returns 'qty'. The real API renamed the field to 'quantity'. All tests passed, and production broke. What would have prevented this?

  37. 37.A team's tests use in-memory SQLite, but production runs Postgres. What is the main risk?

  38. 38.When is it most worth running tests against a real database instead of a fake?

  39. 39.A unit test for a loan-approval class mocks one of that same class's own private helper methods. What is the problem?

  40. 40.A partner's shipping API allows only 100 calls per day in its test environment, and your suite makes 2,000. What is the best approach?

  41. 41.What is a flaky test?

  42. 42.A browser test clicks Save and at once asserts that a 'Saved' toast appears. It fails about 1 run in 10 on the build server. What is the most likely cause?

  43. 43.A browser test races an asynchronous save before checking for a confirmation message. What is the best fix?

  44. 44.A test passes when the full suite runs but fails when run on its own. What is the most likely cause?

  45. 45.How do you surface hidden order dependence across a large test suite?

  46. 46.A team configures its pipeline to retry every failing test up to three times. What is the main risk?

  47. 47.A flaky test in the merge gate blocks the whole team several times a day. What do you do first?

  48. 48.Tests run in parallel workers, and all of them use the same message queue name. Tests sometimes read messages that belong to another test. What is the best fix?

  49. 49.A pet groomer app's test for 'tomorrow's appointments' fails only between 7 p.m. and midnight Eastern time on the build server. What is the most likely cause?

  50. 50.A theater seat-booking test fails in about 2 percent of runs. Logs show two concurrent requests were both given the same seat. What do you do?

  51. 51.What does test isolation mean?

  52. 52.A full regression suite takes 6 hours, and pull requests wait on it. What is the best approach?

  53. 53.What is a known weakness of choosing which tests to run based on which files changed?

  54. 54.A bug caused rental deposits to be refunded twice. It has been fixed. What should be added?

  55. 55.A suite has 4,000 tests and is slow, and many seem redundant. How do you trim it safely?

  56. 56.Several unit tests change a shared singleton config object, and failures appear at random. What is the best fix?

  57. 57.What is a smoke test suite for?

  58. 58.Which description best fits exploratory testing?

  59. 59.Which is the best charter for an exploratory testing session?

  60. 60.A field-inspection app adds an offline mode, and the specs are still vague. How do you start testing it?

  61. 61.What should you record after an exploratory testing session?

  62. 62.You are exploring a free-text 'notes' field in a vet clinic app. Which set of inputs is most likely to find problems?

  63. 63.For which feature is exploratory testing more valuable than adding automated checks?

  64. 64.What is the difference between severity and priority for a defect?

  65. 65.Just before launch, the company name is misspelled in the homepage header. Nothing else is affected. How would you classify it?

  66. 66.A crash occurs when exporting to a legacy format used by two internal staff, who have an easy workaround. How would you classify it?

  67. 67.Which details matter most for a developer to reproduce a bug from your report?

  68. 68.Which is the best title for a bug report?

  69. 69.A developer marks your bug 'cannot reproduce'. You can still see it. What do you do next?

  70. 70.While testing one page, you find three unrelated problems. How should you report them?

  71. 71.A bug in a tram ticket kiosk appears about once in 20 tries. What should the report include?

  72. 72.What are exit criteria in testing?

  73. 73.All planned tests pass, but one open high-severity bug breaks payment for one card type. What do you do?

  74. 74.Which is the best exit criterion for a release?

  75. 75.Someone asks you, as the test lead, 'Is it ready to ship?' What is your role in that answer?

  76. 76.A risky change to a food bank's donation-matching logic ships on Friday. Besides tests, which readiness check matters most?

  77. 77.Which signal suggests a build is not ready, even though the planned tests are done?

  78. 78.How should a continuous integration (CI) pipeline order its checks to give developers the fastest useful feedback?

  79. 79.Tests pass on developers' laptops but fail in the continuous integration (CI) pipeline. What is the most likely kind of cause?

  80. 80.What does 'shift-left' mean in testing?

  81. 81.You join a story refinement meeting for a new feature. What is the most valuable thing you can contribute?

  82. 82.Developers hand code to testers at the end of each sprint, and testing has become a bottleneck. Which change best reflects shift-left?

  83. 83.Why spin up a fresh, short-lived environment for each pull request's integration tests?

  84. 84.The main branch has failed its pipeline for two days, and people keep merging on top of it. What do you push for?

  85. 85.What is the difference between a load test and a stress test?

  86. 86.A concert ticketing API reports an average latency of 120 ms, but users complain it is slow. What do you check?

  87. 87.A load test shows an app failing at 50 users, yet production handles 2,000 users fine. What is the most likely explanation?

  88. 88.What makes a load test scenario realistic?

  89. 89.A soak test runs a system at normal load for 12 hours. What kind of problem is it most likely to find?

  90. 90.A search page in a used-car marketplace feels slow. Before tuning anything, what do you do first?

  91. 91.Under load, latency rises while app CPU sits at 20 percent, and many requests are waiting to acquire a database connection. What is the likely bottleneck?

  92. 92.Which quick manual check finds many accessibility problems in a sign-up form?

  93. 93.An automated accessibility scan of a city parks website reports zero issues. What does this mean?

  94. 94.A screen reader announces 'image, image, image' on a recipe page. What is the defect?

  95. 95.A credit union's app shows light gray footer text on a white background. Which check applies?

  96. 96.When a modal dialog opens, keyboard focus stays on the page behind it. What should the fix ensure?

  97. 97.A form marks invalid fields only with a red outline. What is the accessibility problem?

  98. 98.A POST /bookings request is missing the required date field. What should a good API test expect?

  99. 99.A payments API supports idempotency keys. Which test checks this feature?

  100. 100.Which API test checks for broken object-level authorization?

  101. 101.An API test checks only that the status is 200 and one field has the right value. What is a sensible addition?

  102. 102.A list endpoint returns orders in pages of 50. Which test is most likely to find a real pagination bug?

  103. 103.Beyond the happy path, which set of tests is most useful for a /transfers endpoint?

  104. 104.What does 100 percent line coverage tell you?

  105. 105.A service has 90 percent coverage, yet many of its tests contain no assertions. What does this show?

  106. 106.Code contains if (isMember && hasCoupon). Tests only run cases where both are true. What does branch coverage reveal that line coverage might not?

  107. 107.What does mutation testing measure?

  108. 108.A manager sets a hard 95 percent coverage target for every team. What is the main risk?

  109. 109.What is the best use of a code coverage report?