We've had some good results from using

async

methods to run independent queries. One use case is when the single query might be performatively expensive due to the number of joins and we're able to refactor it into discrete independent units that we can merge afterwards. There are also cases where, when querying large sets of transactions (like for inventory balance purposes), we'll break the query into chunks by year or quarter, run them through a Promise.all, then combine after the fact. Part of that is due to better performance with smaller result sets with fewer pages.

Simpler searches tend to run faster independent of synchronicity 🙂