# Service Boundaries

{{< auto-cards >}}


Dedicated Hypervisors operates with the following constraints:

Zone-dependent availability
: Not all hardware generations and configurations are available in every zone.

GPU availability constraints
: GPU-enabled dedicated hypervisors are only available in selected zones and may be subject to capacity constraints.

No underlying host abstraction
: You are responsible for managing resource allocation across your instances on the hypervisor.

Maintenance impact
: Certain maintenance operations require host-level intervention, which can affect all instances running on that hypervisor. To ensure zero downtime during patching or maintenance, we recommend provisioning Dedicated Hypervisors with at least N+1 redundancy. This ensures you have a minimum of one spare host available to absorb migrating instances.