Every application deploys to eu-central-1 (Frankfurt), in its own environment. It's the only region available today.
Deploys are not zero-downtime: every deploy briefly interrupts the running application, and recovering from a bad one means redeploying an earlier commit, not a hot restore.
That's deliberate, for now. This is a development platform, not a production host: the place where consultants, agencies and internal teams build and prove things out. Scalable environments are part of where we're headed. Today, when a specific application needs production-grade scale (autoscaling, high availability, a CDN), you host that piece elsewhere. Development keeps happening here.