This would be really useful when debugging mass indexing and trying to find out whether bulk works are big (1000 objects in the payload, meaning 500 bulked works) or small (2 objects in the payload, meaning only one bulked work).