We need some sort of alerting outside of Harness when an autostopping rule fails to restart an instance as scheduled. Harness shows logs when an instance fails to restart, but we need an alert to bring attention to it