An inside have a look at how OpenAI scaled PostgreSQL to tens of millions of queries per second utilizing replicas, caching, fee limiting, and workload isolation. Source link
