Since around 2026-09-08, I can’t open the hosted Grafana at fly-metrics.net. The page never loads — in a normal window, in a private window, and from two different networks.
The Grafana process itself looks healthy; it’s the /login route specifically that never responds.
What I see
| Path | Result |
|---|---|
/ |
302 → /login |
/api/health |
200 — {"database":"ok","version":"11.2.2","commit":"6b0cddaf..."} |
/login |
times out, 0 bytes received (reproduced at 20s and 60s timeouts) |
/login/fly |
302 → /login |
/favicon.ico, /public/build/index.html |
302 → /login |
So every entry point funnels into /login, and /login never responds. /login/fly redirects back to /login instead of starting the OAuth handoff.
Reproduction
$ curl -s -o /dev/null -w "%{http_code} %{time_total}s %{size_download}b\n" --max-time 60 https://fly-metrics.net/login
000 60.001554s 0b
curl: (28) Operation timed out after 60001 milliseconds with 0 bytes received
$ curl -s https://fly-metrics.net/api/health
{"database":"ok","version":"11.2.2","commit":"6b0cddaf9cdbee0b321aa2c51f586d380bbbb3b7"}
Consistent: 3/3 attempts, no session/cookies involved. `fly-metrics.net` resolves to `168.220.81.101` from here.
Consistent: 3/3 attempts, no session/cookies involved. fly-metrics.net resolves to 168.220.81.101 from here.
What I ruled out
- Not a stale session — happens in a private window and with plain
curl, no cookies at all. - Not my network — reproduced from two different networks/machines.
- Not the metrics backend — the org Prometheus API
(api.fly.io/prometheus/``<org>/api/v1/...) answers normally and returns real data, so only the Grafana UI is affected. - The status page shows “All Systems Operational” and Metrics at 99.96%, which doesn’t match what I’m seeing.
Anyone else seeing this? Is there another way into the dashboards while /login is down?
