Understand Retry Policies

Retry policies define max attempts and backoff timing. The default policy uses eight attempts over roughly one day.

Custom policies should account for destination recovery time and duplicate side-effect risk. Each attempt is recorded in the event timeline with response status, duration, and error details when available.

HTTP 2xx responses are treated as successful deliveries. Other HTTP statuses, timeouts, DNS failures, TLS failures, and connection errors are eligible for retry until the configured attempt limit is exhausted. Exhausted jobs move to the dead-letter queue for review.