Test detail
One test across its runs - its rates, its median duration, its runs per day, the errors it failed with and every run with its result.
Open a test from the Test explorer. The page shows the test’s name, its file with its line and browser, and the annotations of the test’s latest run - whatever period is picked. Open in repository opens that line at the latest run’s commit, the same link as on Test in a run. The Test explorer link above leads back to the list.
Summary and filters
Section titled “Summary and filters”One sentence under the name says how the test did in the period, for example “Ran in 42 CI runs during the last 30 days: 3 failed, 8 were flaky and 31 passed first try.” When it ran on more than one branch, the sentence says so (“…, on 2 branches: …”).
Under it, the filters read as a sentence and apply to the whole page: “Showing CI and local runs on all branches from the last 30 days”. Each choice opens a menu:
- Source - “CI and local runs” by default, “CI runs” or “local runs”.
- Branch - “all branches” by default, or one branch the test ran on.
- Period - the last 7, 14, 30 or 90 days; 30 by default.
A test opened from the list keeps the list’s source, branch and period. The ⓘ after the sentence says which runs are counted, for example “Completed runs from the last 30 days on all branches …” - “Completed CI runs …” under CI runs.
| Tile | Shows | Under the value |
|---|---|---|
| Stability | runs that passed on the first try | “16 of 20 first try” |
| Flaky rate | runs that passed only after a retry | “3 after a retry” |
| Failure rate | runs whose last attempt failed | “1 of 20 failed” |
| Median duration | the median length of the last attempt in each run, skipped runs left out | the median of the period before, for example “5.2s the 30 days before” |
| Time lost | retries, and the attempts of failed runs | “retries + failures” |
They count the test’s runs of the chosen source - a sharded run once, when all its shards have finished - like the Test explorer; see How Squally measures tests. Only bad values are red: stability below 70 % and a failure rate of 20 % or more.
Runs per day
Section titled “Runs per day”A bar per day of the period, oldest on the left, in the same runs the tiles count. Each bar is stacked: failed at the bottom, then flaky, then passed. A day without runs is a short mark on the baseline, not a zero bar. The legend gives each part’s total over the period, for example “Failed 3”, and “no runs” when a day had none. Hover a day for its numbers. On a phone the chart scrolls sideways and opens on the newest days.
Errors
Section titled “Errors”The errors this test failed with in the period’s counted runs, one row each, like the Error explorer: the error’s title, its file and line (or how many files it spans) - led by the kind when the title is the test’s own assertion message - in how many runs it occurred, how often those runs failed or were flaky (“failed 2× · flaky 1×”), when it was last seen, and its category. Each opens the error’s page; All errors opens the error list. When earlier failed runs of the test fall outside the period, the card says how many.
Every run of the test in the period, newest first: its result, run number, branch and retries, each in its own column, then the duration with how long ago below it. On a phone the branch and retries move onto a grey line under the run number. Local runs and runs that have not finished are listed too, whatever the source, so you can find them; local runs are marked Local. Whether a run counts in the tiles depends on the source.
When the runs ended differently, segments filter them by result - All, then only the results that occurred (Passed, Flaky, Failed, Skipped), each with its count. When every run ended the same, one sentence says so instead, for example “All 40 runs failed.” Search by run number.
The list shows 20 runs; Show 20 more adds the next 20. A run opens the test in that run.
A test with no runs in the period
Section titled “A test with no runs in the period”The rates show “—”, and the page says “No runs of this test in the last N days”, with a link to the next longer period.