App deploy and routes

A FleetApp deployment proceeds through these stages:

1. An authenticated operator publishes a chart to the library (FleetChart) or references an existing OCI chart.
2. The operator creates a FleetApp with required clusterName and chart onto a Ready cluster (capability filters are optional helpers only).
3. The fleet-site controller runs the Helm install job on the spoke.
4. When route.mode is hub or site, fleet-edge reconciles a FleetAppRoute.

none leaves ingress to the workload and does not create a FleetAppRoute. route.mode: dual is not supported in v0.2; use hub or site separately.

Mode	Intended use	Hub required for public URL?
hub	Hub gateway forwards to the selected workload cluster	Yes
site	Bound site ingress (port_forward / reverse_proxy) exposes the app	No
none	No fleet-managed public route	No

Match route.mode to site edge: hub with hubGateway; site with a cluster ingressId bound to port_forward or reverse_proxy.

Hub-routed apps stop when the hub is unavailable. site and none apps can remain reachable when site ingress is healthy.

Deleting a FleetApp removes the Helm release and associated route. Verify workload removal, routes, credentials, and data handling separately.

See FleetApp for publish, install, and verification detail.