Receiver management improvements for the admin Receivers tab:
- Per-receiver health: server now tracks each connection's state
(connecting/connected/down), message count, and last-message time, exposed
via new GET /api/admin/receiver-status and POST /api/admin/test-receivers.
- Receivers tab redesign: live "Active Receivers" panel with colored
indicators (green=receiving, yellow=connected/no data, red=unreachable),
a clear select -> SAVE & APPLY -> auto-test flow, and selectable scan
results with a "USE SELECTED RECEIVERS" action.
- Network scan interface selector: scan a chosen interface/subnet instead of
every local subnet. Oversized subnets (>/20, e.g. a docker /16) are disabled
in the UI and rejected server-side, eliminating slow/hanging scans.
Critical fix — server crash on receiver restart:
- Receiver connection tasks were awaited inside main()'s top-level
asyncio.gather. Restarting receivers cancels those tasks, and the resulting
CancelledError propagated into the gather and killed the whole server
process (every action after a restart then failed). Receiver tasks now run
as independent background tasks; restart cancels and awaits them with
return_exceptions=True so cancellation can never tear down the server.
Also cache-bust admin.css/admin.js/pixel-editor.js (?v=) so updated assets
load without a manual hard-refresh, and surface a clear message instead of a
silent "Loading…" when the status endpoint is unavailable.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
root
7a80057cb5