We’ve engineered Vercel so that live deployments are unaffected by any gap of availability or performance of our core databases. How? We employ a multi-tiered caching architecture and a global metadata replication system. But: what if the database or core region that manages deployments, rollbacks, new projects went fully down? What if our core region suffered from a natural disaster? At the limit, not being able to revert a bad deployment can be just as damaging as dropping live CDN traffic on the workload. We ran a successful simulation of this disaster scenario and wrote up on it ↓
We recently ran a successful live failover of our control plane database from Azure West US to East US 2. Our goal was to validate that Vercel can survive a complete regional failure of its core database. Here's the full breakdown: vercel.com/blog/preparing…