How many folks actually use multi-region deployments with automatic failover (e.g. latency based routing in route 53)?
back
1 comments
The difficulty of making this work in practice is pretty high. It also isn't cheap, so I would guess not many.
Some parts are expensive (EC2, NAT and VPC endpoint proliferation) but others are simple and inexpensive (regional API Gateways, latency based routing, DDB global tables, lambdas, state machines).