Run Failed
A failed run means a test failed or the run could not execute successfully, including failures before any scenario started. It does not always mean the application is broken. The failure may come from the app, the test, the target data, or the environment the cloud test session ran against.
First Review
- Open the run detail page.
- In Tests, filter to Failed. If no scenario failed, read the run-level error and check skipped or unreported rows.
- Open the failed row or scenario.
- Read the error message.
- View the screenshot or video near the failed step.
- Open console, network, or trace evidence if needed. Debug with Assistant can investigate and prepare a supported fix for review.
What Failed?
| Evidence | Possible meaning |
|---|---|
| Screenshot shows the wrong page | The test started at the wrong URL, login failed, or navigation changed. |
| Element not found | The app UI changed, the selector is unstable, or the step ran too early. |
| Timeout | The app was slow, the session could not reach the site, or the test waited for the wrong thing. |
| Console error | The application may have thrown a browser-side error. |
| Network error | An API call failed, was blocked, or returned unexpected data. |
| No tests executed | The selected test, dependency, or run setup may be wrong. |
Decide The Next Action
| If the evidence shows | Do this |
|---|---|
| The application behavior is wrong | Fix the app or test environment, then re-run. |
| The test expects old behavior | Edit the feature or generated code. |
| The locator or timing is weak | Regenerate or edit the Playwright test. |
| Test data is missing | Add explicit setup steps, secrets, or test data. |
| The session cannot reach the site | Check the site URL; for private apps, check the tunnel device. |
Before Re-Running
Re-run after you make a meaningful change. Re-running without changing anything is useful only when the failure looks temporary, such as a network timeout or a staging outage.
Choosing Re-run on the run opens the run composer pre-filled with that run's tests and configurations, targeting the same site URL — nothing starts until you choose Start run or Start Assistant run. Concurrency and retries use the current project defaults; review Run options if needed. Use the composer to narrow the retry to just the failed tests instead of repeating the whole run. See Cancel or re-run tests.