Tagged
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.
Bloom filters shave 90% of negative lookups; HyperLogLog estimates cardinality without storing every user ID. Redis modules on ElastiCache for abuse detection and feed deduplication.
Cache-aside without an invalidation story ships stale pricing to 2% of users—the hardest 2% to debug. This guide layers CloudFront, ElastiCache, and DAX with TTL, event-driven purge, and when write-through beats cache-aside.
Redlock debates matter because ElastiCache is not a consensus system. Consistent hashing for sharding workers and ALB target stickiness—with DynamoDB conditional writes as the boring alternative.
RDS performance: gp3 IOPS, Performance Insights / Database Insights, Proxy, replicas, ElastiCache. July 2026 tuning checklist.
Savings Plans and Reserved Instances reduce the rate you pay. Architecture determines the volume you pay at. The most durable cost reductions in AWS come from designing systems that structurally generate less spend — not from negotiating a lower price for the same behavior.
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.
Valkey on ElastiCache: about 20% node / 33% serverless cheaper than Redis OSS engines. July 2026 — sessions, rate limits, locks, migration checklist.
Redis is fast — until your application retries on every cache miss and the Redis bill starts looking like the database bill. ElastiCache patterns, data structures, cluster modes, eviction policies, and the production patterns that actually reduce database load.