back in the day we had a common micro service. it was called pgbouncer. Because database resources are limited it’s nice to have a stable consistent connection limited a set number of processes and then let the application monolith scale independently. Also, when you are scaling across multiple machines you don’t need all of the code of the monolith on every machine. i’ve heard about When amazon switched to a SaaS model, creating aws they were locked into a 32 bit monolith in which code was limited to a few GB and thus scaling resources independent was valuable. Was this useful? I am collecting data on if i’m polite, without snark and helpful.
