Service Status page reports live health across all deployments

Fixed the admin Service Status page reporting a healthy gateway as down. The page probed a health endpoint the web proxy did not forward, so the check failed on every deployment, and the status table was hardcoded for Docker and wrong on Kubernetes.

Status is now served by a platform-aware endpoint: the gateway checks its database, cache, companion services, model server, and monitoring stack server-side over its own configured URLs and returns a row per service. A down dependency shows as a row, never an error page, and only the exception class is surfaced — no internal detail.

No action required.