Read a failed run

Find which step failed and why, compare it with the last pass and fix runs that fail for outside reasons.

In short

A failed run opens on the step that failed, with its screenshot, the error and the selector it used. It sets that step beside the same step in the last passing run, so you see what changed.

Before you start

• Who: anyone who can open the site reads its runs (View).
• Plan: run screenshots and videos are kept for 30 days; the runs and their verdicts stay.
• First: open a failed run.

Read the failure

1. Read the banner: "Element not found", "Assertion failed", "Timed out", "Navigation failed" or "Browser error", at step n.
2. Open the failed step: the error, the selector it used, the page address and a slice of the page around where it looked.
3. Read the screenshots side by side with the last passing run, and what else differs under them.
4. Press "Why did this fail?" for the assistant's explanation, when AI is on for you.

Compare with the last pass

A failed or healed step shows its screenshot beside the same step in the last run of the same test version that passed. Steps are matched by the step itself, so reordering does not pair the wrong ones. The panel names the selector that found the element, or says that nothing did. It also names the page and any step that became much slower, and says when the passing run used another environment.

Good to know

• Steps after a failure are Skipped. An optional step may fail without failing the run.
• "Run details" lists the source, environment, times, browser, viewport, test version and parameters, with secrets as *.
• The last passing run's screenshots are kept even after older ones are cleared.
• A sensitive value typed in the run options is held only until the run starts and is never stored.

If something goes wrong

Related

• /help/ui-tester/review-steps
• /help/ai/why-did-this-fail
• /help/home/needs-you