Runs: Starting, Progress, and Errors
A run is one execution of a scenario. It has a selected version, parameters, state, log, and results. Starting a run creates a task; execution may begin later.
States and actions
| What you see | What it means | What to do |
|---|---|---|
| Queued | The task was accepted, but a worker has not started | Check the agent and other runs |
| Running | Collection and processing are in progress | Watch stages, logs, and pages |
| Completed | Processing has finished | Check records and quality |
| Failed | Execution has stopped | Read the reason and the latest messages |
| Canceled | A user stopped the run | Decide whether to start a new run |
Run page
The overview shows duration, stages, and custom scenario steps. Data updates while the run is active. The percentage may be approximate if the total amount of work is still being determined. Completing one stage does not necessarily mean the whole run is complete.
In the log, find the first meaningful error message, not only the final status. In the page list, check the URL, response code, and duration. A snapshot opens in the right panel and shows the content received at the time, which may differ from the page you see now in your browser.
If a run remains queued
Check the selected worker, its availability, concurrency limits, schedules, and tasks already running. Do not click Start repeatedly: each accepted request may create a separate run.
After an error
Fix the cause before retrying. A limit refusal calls for a budget check. A source HTTP error calls for a check of the address, availability, proxy, or rendering. A data transformation error calls for a check of the scenario and fields. Whether a run can resume depends on its state. A new run may request the source again and consume your limit.