Skip to main content
Idempotency keeps a workflow from running more than once for the same logical event. When you fire an event for a linked account, a retry or a double-send can deliver that event twice. With idempotency on, Refold reads a key from each incoming call and ignores any repeat that carries a key it has already seen inside the configured window. Use it on event-triggered workflows where the caller might send the same event more than once, so a duplicate delivery never creates a duplicate run.

​
How it works

You configure idempotency on the workflow’s event trigger. Refold extracts a value from each incoming event and treats it as the call’s unique identity. If another call arrives with the same value before the time period elapses, Refold drops it instead of starting a new run.
SettingWhat it does
Key fromWhere Refold reads the key from on the incoming event: the request Header or the Payload.
Idempotency keyThe path to the unique value, written as {event:<key>}. The value at {event:<key>} uniquely identifies each call in the selected source. For example, {event:id} uses the event’s id.
Time periodHow long a key is remembered. Duplicate calls within this window are ignored. Defaults to 5 minutes.

​
Configure idempotency

1

Open the trigger settings

In the workflow, open the event trigger and find the idempotency settings.
2

Choose where the key comes from

Set Key from to Header or Payload, depending on where your caller puts the unique value.
3

Point to the unique value

Set the Idempotency key to {event:<key>}, where <key> is the field that uniquely identifies each call, for example {event:id}.
4

Set the time period

Choose how long duplicates are ignored. The default is 5 minutes.
Pick a key that is stable across retries but unique per logical event, such as an upstream record ID or event ID. A value that changes on every send (like a timestamp) defeats deduplication.

​
See also