bumping this one as its a new day and has fallen off, is there any way to fix this short of re-indexing all of the data from postgres into a fresh cluster?