What a task is

Tasks are background jobs that turn provider downloads or uploads into project assets, process models, and perform other queued preparation. You can leave the page while a task runs. Its progress remains available from Tasks and the header notification center.

A queued task is waiting for capacity. Running is active work. Retrying means an automatic or requested retry is pending. Completed means the result is ready. Failed means the latest attempt ended with an error. Archived hides finished history from the default active view; it does not undo the produced asset.

Find the task you need

Open Tasks from the Overview sidebar. The first filter shows Active only, Archived only, or all history. The second shows All statuses or narrows the list to Queued, Running, Retrying, Completed, or Failed. Search uses the Task ID, project, provider placeholder.

Check the active team before searching. A task belongs to the team that owns the project or model, so it will not appear under another selected team.

Inspect task details

Select a task row to open its details. The summary shows the project when applicable, creation and update times, priority, retry count, archive state, requested and current provider, and usage. Flags explain special choices such as upload mode or fallback.

While the job runs, watch the progress and live log. On completion, use the result action to open or download the produced asset or backup when one exists. On failure, read the final error and the last useful log lines before retrying; they usually identify credentials, provider coverage, input format, worker capacity, or processing failure.

Stop, archive, retry, and use bulk actions

Select one or more queued, running, or retrying tasks and choose Stop and archive when the work is no longer wanted. Stopping asks the worker to end the job and removes it from the active list after the state is saved.

Select completed or failed tasks and choose Archive to clean up the active list without deleting their project results. Select a failed task and choose Retry after correcting the cause. A retry consumes capacity again and may select another available provider or worker when the task allows it.

Bulk actions apply to every selected row that supports the action. Clear the selection and filter by status first if the list mixes jobs you intend to keep and jobs you intend to stop.

Task units and waiting work

A task normally occupies one hourly task unit while it is active. The unit returns after the rolling hour shown on Usage, so finishing a task does not always make that unit immediately available. Some optional extensions can add capacity.

A task can also wait because no compatible worker is online, the chosen provider is temporarily unavailable, required credentials are missing, or the team reached another operational limit. Open the task details first, then check Workers, Data Providers, and Usage according to the message shown.