Execution Retry and Replay

Because external systems are not always available, it is important to implement retry functionality in your integrations where appropriate. Retry enables the system to automatically resend payloads to destination systems at regular intervals or using exponential backoff (e.g., retrying after 2 minutes, then 4, then 8, etc.). Implementing robust retry logic helps ensure your integration succeeds even if the destination system experiences a temporary outage.

By leveraging retry, you can avoid involving your development team in integration issues until you have confirmed the problem is more serious than a transient outage. This prevents many ephemeral issues from reaching your support team.

However, if errors persist (for example, if the destination system appears operational but the integration continues to return internal server errors), you may need to escalate from retry to replay functionality.

Unlike retry, replay is a manual operation. It is initiated by clicking the replay button in the instance UI. Replay reruns the entire integration with the original payload, allowing you to debug by monitoring each step's execution and pinpoint where the process succeeds or fails.