A team went active-active across two regions for a service whose SLA is only 99.95% availability — a level a single, well-built region hits comfortably. The result is double the infrastructure cost and a budget blown wide open.
Targets: 99.95% availability, p99 ≤ 140 ms, critical durability, under $1,200/mo.
Bring it back to a single region done right — two AZs, Multi-AZ database, a cache, Graviton on a Savings Plan — and only pay the multi-region premium if the SLA actually demanded it (here it doesn't).