Skip to main content
A Delivery is one event sent to one endpoint. Elapse POSTs the signed body, waits up to 10 seconds, and counts any 2xx as success. Anything else, including a timeout, is a failed attempt and the schedule below applies.

​
Schedule

After attempt 8 fails, the Delivery is marked exhausted and the dashboard shows Resend. The schedule is read from the worker’s own configuration at build time, so this table is what runs.

​
Respond fast, work later

Answer 200 as soon as the signature verifies, then do your work. A handler that provisions a GPU before responding will time out under load and be retried, and the retry will provision it again.

​
Idempotency

A retry delivers the same event id. Keep the ids you have handled and treat a repeat as a no-op:

​
Failing endpoints

An endpoint that fails continuously for three days is disabled and you are emailed. Re-enable it from the dashboard once it is fixed; nothing is lost, every Delivery is still there with its attempts.

​
The dashboard

Developers → Webhooks shows each endpoint, its success rate, and every Delivery with its attempts: status code, duration, the request headers that were sent. Resend on a Delivery queues one more attempt with the same body and a fresh signature. The same view exists for the CLI endpoint that elapse listen creates, so you can see what your laptop answered.