UptimeUptime Wiki

Geographic regions

A region is a geographic area — US East, EU West, APAC. Customers pay for snappy service by being served from a region near them, so spreading out widens the market you can reach. Some customers go further: banks and healthcare firms are legally required to keep their data inside a specific country, and they simply won't sign unless you operate where their rules say their data has to stay.

Each new region is not a quick purchase — it's a big up-front build, far pricier than adding another building inside a region you already run. More regions means a bigger addressable market and edge presence closer to users, but you're committing real capital every time you plant a flag somewhere new. Expand because the demand is there, not because the map looks empty.

Detailed explanation

Region as a container of AZs

A region holds N availability zones. Cross-region latency runs 60-200 ms, and there is no inter-region failover below that line — a region is its own island for reachability. New-region capex is about 5x a Hyperscale-AZ build, which is why a region launch is a deliberate strategic move, not opportunistic capacity-bumping.

Residency and blast radius

Customers may set required_region, pre-checked at accept_prospect so an out-of-region prospect is hard-rejected at onboarding rather than silently mis-placed. Region-level failures bring every AZ in that region down for 12-72h — the widest blast radius in the model. CDN POPs scale with region count, so footprint and edge reach grow together.

On this page