You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Route reachability runs in the background and alerts. Traefik Manager checks every enabled HTTP route on the Host and on each agent, every 5 minutes by default, and sends a notification when a backend stops answering and again when it is back. Where the service has a Traefik health check that result is used; otherwise the route is requested through the proxy, with the backend tried directly on 502, 503 or 504. Two failed checks in a row before a route counts as down. Toggle and interval in Settings - Notifications
Wildcard routes are checked at the link you set on the Dashboard. A rule like Host(`*.example.com`) has nothing to request, so those routes stayed grey; a link set in the route's Dashboard edit is now used for the check, and without one the tooltip says what to do (#165)
A dead server in a pool reads as degraded. Without a Traefik health check the proxy silently skips a dead server, so a two-server service looked healthy. Each server is now checked directly and the tooltip names the one that is down (#165)
Routes behind forward auth are checked at the backend. Authentik, Authelia and GateKeeper answer with a redirect before the backend is reached, so a ping through the proxy said online for an app that was off. A redirect to another host now sends the check to the backend address; if Traefik Manager cannot reach that address the tooltip says the backend was not verified instead of guessing (#165)
Fixes
A route with a dead backend showed a green dot on the Routes tab. The dot was painted from Traefik's router status, which only says the router loaded, and a manual ping result was thrown away on the next redraw. The dot now follows the background check, and a ping result stays until a newer check replaces it
The stat panel called a dead route healthy. The HTTP routers card, the verdict line and the entry point rows counted a router as live whenever Traefik had loaded it. They now count routes the background check found down as unreachable, and the count opens the Routes tab filtered to them
selfh.st icons were months out of date. The CDN path Traefik Manager loaded them from had stopped refreshing, so a slug returned an old icon while the library had a new one. Icons now load from the branch the CDN keeps current (#165)
Every open browser tab pinged every route. The dashboard and the Routes tab now read the background result instead of each running their own checks