We would like to see a schedule type that can run only errors in the last extraction. The reason being that we might have errors due to timeouts, dropped network connections etc and we would like to automate these reruns.