Krystie
81f6049ded
The dashboard polls /api/v1/services/status (not /probe/:id) for its refresh loop. routes/services.js's requestStatusCode() didn't set the X-DashCaddy-HealthCheck: 1 marker, so the batch endpoint hit the forward_auth gate, got rate-limited by authLimiter (429), and reported 7 services (router, chat, sync, torrent, sonarr, radarr, prowlarr, requests) as down. Same fix in src/app.js /probe/:id (the single-service endpoint) for consistency. Without the marker, every probe from the container IP trips authLimiter within 20 requests and the rest of the batch fails. health-checker.js background poll was already setting the marker correctly, which is why the cached health view showed 15/15 while the live dashboard showed 8/15.