Using your workspace
Routines and webhooks
Run a repeatable agent task manually, on a schedule or after an external webhook event.
Prove the task manually first
Write a small task with a defined input, time period and output. Run it manually and inspect the result before adding a trigger. A schedule starts work; it does not prove that the task completed or that its result is correct.
Set a schedule and check the run
Choose a schedule in the routine settings and check the next run time shown by the app. Review the first scheduled result and its run record. Scheduled execution depends on the service that runs the routine scheduler being available.
Use a webhook carefully
A webhook URL can trigger its routine, so treat it like a secret. Store it in the calling service’s secret settings and rotate it if it is exposed. Configure the caller to retry temporary failures. A request can receive HTTP 503 while a hosted sandbox is waking; the request is not queued for later delivery.
Grant only the access the routine needs
If a routine uses a connected app, grant it the exact account already authorized for that agent. Review the grant and the routine’s output after each change. Keep important scheduled workflows observable and retain a copy of outputs you need.
Need help with a setup? Contact the Suparagent team.