Keep local terminals usable during cloud outages

[fkesheh]

When Fly/Centrifugo becomes unreachable, the dashboard currently blocks access to terminals even when their Local/P2P connection remains healthy. This change keeps the dashboard usable and lets known local bridges reconnect during cloud discovery failures.

• Replace the automatic blocking error panel with a reconnecting notice and dismissible diagnostics.
• Show Local/P2P connectivity separately from cloud status in the main indicator.
• Remember authenticated loopback advertisements in memory for retries when presence rejects, hangs, or loses the bridge; refresh advertisements when discovery returns and clear them on disposal.
• Preserve existing automatic cloud reconnection and healthy direct terminal connections.

Scope: an already-open authenticated dashboard. Offline page reloads, newly discovered/restarted bridges with changed credentials, and session-management RPCs still require cloud access. No bridge protocol changes or deployment are included.

Validation: all 248 UI tests pass on the latest main, including outage/recovery coverage; npx tsc --noEmit and git diff --check pass. Network-outage scenarios use mocked transports; no live outage test was performed.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
ftown	Ready	Preview	Sep 16, 2026 5:50pm UTC