A delivery that does not get a 2xx answer is retried on a fixed schedule: after 30 seconds, 2 minutes, 10 minutes, 1 hour and
6 hours. After the last try the delivery is marked failed. HTTP 410 switches the endpoint off for good; a 429 with Retry-After
delays the next try by that long, up to five minutes. After 20 failed deliveries in a row the endpoint is switched off and the
owner is told.
Waits before each try:
- Try 2: 30 seconds
- Try 3: 2 minutes
- Try 4: 10 minutes
- Try 5: 1 hour
- Try 6: 6 hours
Every retry carries the same X-RankPilot-Delivery id, so your receiver can ignore a delivery it has already handled. The
delivery log under Alerting setup, History shows each attempt (request and response, truncated to 2 KB). Replay sends a
stored delivery again with a new timestamp and the same delivery id; deliveries older than 30 days cannot be replayed, and a
workspace can replay 10 per minute.