Verdicts
| Outcome | Meaning |
|---|---|
| Passed | Connect played what the test expected, and the call ended the way it predicted. |
| Passed on re-run | It failed once and passed on a second run: most likely Connect missed a prompt the first time. |
| Failed | Connect played something else, or the call went elsewhere. Look at the step the test stopped at. |
| Failed twice | It failed on its re-run as well, so it is unlikely to be chance. |
| Not a verdict | Connect could not run the test (capacity, throttling, a service error). It is neither a pass nor a fail. |
| Checked, not run | A dry run: the test was converted to Connect’s format and checked, but nothing ran. |
| Skipped / Stopped | Not run (estimated past Connect’s time limit), or stopped before it finished. |
Inconclusive results (not a verdict, skipped, stopped) are reported separately and never count towards a suite’s pass rate.
Dry runs and live runs
Section titled “Dry runs and live runs”A dry run converts every test and validates it, without creating anything in Connect. Do one
first. A live run creates the test cases in Connect and executes them; Amazon bills these
executions. Live runs need a key with the PUBLISH_LIVE scope.
Interrupted runs
Section titled “Interrupted runs”If the server restarts while an evolution run or a Connect run is in progress, that run is marked Interrupted: it did not finish and was not lost silently. Start it again.