rds

For a fintech SaaS (6 prod databases, ~$14k/mo RDS line), scoping in-house DBA for schema + MSP for 24/7 paging cut on-call from 22 to 9 hrs/mo — Aurora Serverless v2 min ACU tuning handled tier-2 first.

On a composite B2B order API (~2.4k RPS peak, 24 ECS tasks on Aurora), running the scaling ladder in order — RDS Proxy, read-replica routing, ElastiCache, then SQS offload — moved modeled p99 from 1,840 ms to 72 ms before any microservice split.

Too many "too many connections" pages are fixed by raising max_connections—which trades one outage for OOM on the writer. This guide traces deadlocks, pool sizing, RDS Proxy, and prepared statement caching on Aurora.

Serializable sounds safest until your checkout times out under row locks. This guide maps READ COMMITTED, REPEATABLE READ, and SERIALIZABLE to RDS/Aurora defaults—and when DynamoDB conditional writes are the BASE alternative.

A B2B SaaS stack that costs $500/month at launch does not need to cost $50,000/month at 100,000 users if the architecture decisions at each stage are deliberate. This is the end-to-end reference architecture with real cost numbers.

RDS, Aurora, and self-managed Postgres each have a cost breakeven point. This guide covers total cost of ownership, connection pooling with PgBouncer, indexing strategies, and the edge cases that turn Postgres into a billing surprise.