I benchmarked pymongo vs. asyncmongo and found no performance gain (or even negative gain) with the latter unless making relatively heavy/slow queries. So, the asynch spaghetti is not worth of it in most cases, just like the authors experienced with MySQL.