CrowBar | Delivery and retries | Corvus Skip to main content

Delivery and retries

  • Respond with any 2xx status to acknowledge an event. Any other status is treated as a failure. Redirects are not followed.
  • Each request has a 15 second timeout to connect and a 15 second timeout to receive your response.
  • Failed deliveries are retried every 5 minutes, up to 10 attempts in total. After the last attempt, the event is not redelivered.
  • Delivery is at-least-once. Events may arrive out of order, and the same event may arrive more than once.
  • An event your partner is not subscribed to is skipped, not retried. Subscriptions are checked when the webhook is sent, so a change also applies to webhooks that were already queued. See Event subscriptions.
  • Each attempt has a new timestamp and signature, so the timestamp cannot be used to detect a repeated delivery. Some events also legitimately occur more than once for the same resource. Make your handlers idempotent: treat each event as a signal to query the resource for its current state, rather than as the state itself.