Failover

v0.2 disclaimer: failover is optional, incomplete, and must not be treated as a production availability guarantee. Multi-cluster hub routing is experimental active–backup behavior, not active–active load sharing.

When standby routing is configured for hub-routed apps, the fleet-failover controller may update the selected route’s active cluster from hub health signals. Both workload installation, application state, DNS, TLS, firewall behavior, and recovery remain deployment-specific responsibilities.

Per-site port_forward and reverse_proxy exposure is independent of hub-gateway failover.

Validate failure detection, data consistency, failback, and externally reachable traffic in the target environment before representing a route as resilient. Stateless applications are a more realistic target than stateful workloads.

For the current console walkthrough with HA screenshots (captured against v0.1 Catalog), see console walkthrough sections 9–10 with v0.2 disclaimers.