Skip to main content

Webhook retries

Webhook retries recover from temporary downstream failures without asking the provider to resend the original event. trusin records each attempt so an operator can distinguish a transient failure from a final failure.

Retry behavior​

For retryable network failures, trusin changes the event to retrying and schedules exponential backoff: 10 × 2^retry_count seconds. After MAX_RETRIES, the event is marked failed. A successful 2xx response marks it delivered.

Design receivers for retries​

  • Make processing idempotent: store and deduplicate the provider event ID.
  • Return 2xx only when the event is safely accepted.
  • Treat client-side validation failures as final; use an appropriate non-2xx response when the payload cannot be processed.
  • Use the delivery timeline to inspect request and response details before a manual retry.

See the reliability model and troubleshooting guide for operational details.