The cleaner currently prints the 'cleaned size' and the largest known checkpoint size after every run which can be useful to see if the cleaner is keeping up with log growth. We should convert these to prometheus metrics so it can be part of regular operational monitoring without having to check logs.