Schedule an Executable
Start with an uploaded release and active profile. You can configure the schedule in the executable wizard, edit it later, or use the management CLI and API.
Choose a cadence
Use an interval for regular polling:
{ "type": "interval", "intervalMs": 60000 }Use cron for calendar-aligned work:
{ "type": "cron", "cronspec": "0 * * * *" }An interval's optional startAt is a future epoch timestamp in milliseconds that defers registration. Check the console preview and displayed next run rather than assuming your browser timezone determines execution time.
Leave enough time for task reads, submission, and confirmation. A fast interval does not create parallel access to the same persistent state, and external API limits still apply.
Preview, then activate
Configure cloud args and bind secrets, review scope coverage, and use the simulation action to inspect produced calls. Once ready, Execute now requests a real invocation independent of the schedule. Open its logs and verify the expected result.
skipped can mean an intentional task decision or a simulation/gate. Read its message. submitted_unknown means the transaction can still land and should be reconciled before trying to repeat the work.
Update or stop the schedule
Editing the trigger replaces its registration. Pausing prevents new enabled invocations and leaves stored state available. It does not cancel a transaction already submitted. Resume rechecks the release/profile binding and restores the schedule.
Repeated policy or funding enforcement failures can automatically pause work. Inspect the reason, correct the cause, and activate deliberately; do not assume a top-up alone resumes a paused executable.
External events
For off-chain event delivery, create a named webhook URL and use an idempotency key per logical event. The recurring schedule remains active alongside webhook calls.
For on-chain events, poll from the task and carry an appropriate cursor. There is no native event/log trigger in the stored executable trigger model.