hard20 pts · first answer countsdatabasesscalabilityperformance

Connection pools and autoscaling

Each API instance has a PostgreSQL pool of 50 connections. During a traffic spike, autoscaling grows from 4 to 40 instances, and the database starts failing with too many connections and high latency. What is the best fix?

A.Increase every instance pool to 200 connectionsB.Use smaller per-instance pools and a connection pooler like PgBouncer, sized to what the database can actually handleC.Open a new connection per query and close it right afterD.Disable autoscaling permanently

← One slow route freezes Node.js Why a lock alone is not enough →