Skip to content

Triggers and scheduling

A trigger belongs to a process definition and answers two questions: when does a run start, and what is its due date? Flowstate calls the due date the anchor; every activity offset is measured from it.

Type Starts a run when Anchor
Manual You publish the process from the app. The date and time you enter.
Schedule A recurrence rule fires, for example the last working day of every month. The occurrence’s date and time.
Source event A connected source produces an event, for example a calendar entry. A field on the event, such as its start time or a custom due-date field.
Webhook An external system calls the trigger’s URL. Supplied in the request.
API Your own code calls the API. Supplied in the request.

Manual, schedule and source-event triggers are available today. Webhook and API triggers can be defined but cannot start runs yet; they will become active together with the public API.

A process can have more than one trigger. Each run records which trigger started it.

Schedule triggers use a recurrence rule and a timezone. If the service was unavailable when an occurrence was due, the misfire policy decides what happens on recovery: catch up all missed occurrences (up to a cap) or skip to the most recent one.

The anchor’s timezone determines what “09:00” means for every activity in the run. For source-event triggers you choose the strategy:

  • Source event timezone - use the timezone the event carries (a calendar entry created in Oslo stays on Oslo time).
  • Trigger timezone - always use the timezone configured on the trigger.
  • UTC - treat the anchor as UTC.

Manual triggers use the workspace timezone.

Publishing a run is preceded by a plan: Flowstate shows the resolved date and time for every activity before anything is scheduled, so you can catch an offset that lands on a weekend or a public holiday before the emails go out.