Delivery and retries

Network failures, timeouts, 408, 425, 429 and any 5xx are retried. Other 4xx responses are treated as permanent. If you return 400 because you could not parse the body, you will not see that event again.

Retry schedule

Attempt	Sent
1	Immediately
2	1 minute later
3	5 minutes later
4	15 minutes later
5	1 hour later
6	3 hours later
7	6 hours later
8	12 hours later

Eight attempts spread over roughly 22 hours. An endpoint that is down for an ordinary deploy still receives its events.

When the attempts run out

There is no replay

Once the attempts are exhausted, the event is not delivered again. Nothing in the portal can resend it. If your endpoint was down for longer than the retry window, reconcile against the payments list instead of waiting for a redelivery.

Delivery history

The portal shows the ten most recent delivery attempts, including the response code your endpoint returned. This is usually the fastest way to tell a signature bug (401) from a routing bug (404) from a slow handler (a timeout). Delivery history is kept for 90 days.