status.nfltr.xyz
The hosted relay, as the canary sees it
Every hour a canary runs the whole orchestration path against nfltr.xyz: a fresh hub, two nodes joined through the relay, an agent on each, a monitor, every result delivered exactly once and a pushed branch checked on its origin. The agents are a test double, so no model runs. This page shows those runs.
The last canary run passed.
No canary run in the last 2 hours, so these results are out of date. A gap means the canary did not run (its machine was asleep or offline), not that the relay was down.
Summary
- 2026-10-15 01:12 UTC
- start of the newest run
- v2.1.1
- relay version in the newest run that read it
- 100.0 %
- Last 24 hours: 20 of 20 runs passed
- 100.0 %
- Last 7 days: 119 of 119 runs passed
Latency
Measured on the canary's machine from just before each agent is started, over the agents of passing runs. The test double takes 3 s of each completion.
| Start to first progress, p50 | p95 | Start to completion, p50 | p95 | |
|---|---|---|---|---|
| Last 24 hours | 11.7 s | 14.3 s | 17.1 s | 19.9 s |
| Last 7 days | 12.5 s | 22.0 s | 17.4 s | 34.4 s |
Judged after each run: latency is degraded when each of the newest passing runs is well above the median of the passing runs in the days before them.
| Baseline | Newest runs | ||
|---|---|---|---|
| the agent start call | 4.6 s | 2.4 s | normal |
| start to first progress | 12.2 s | 6.2 s | normal |
| start to completion | 17.4 s | 12.5 s | normal |
Recent failures
No failed run in the last 7 days.
How to read it
One machine runs the canary
A gap in the runs means that machine was asleep or offline, not that the relay was down. The page says so once the newest run is more than 2 hours old.
Published after the runs
The canary's machine publishes this page after its runs; this copy was made at 2026-10-15 01:13 UTC. The results stay on that machine; the relay stores nothing for this page.
What it does not cover
Share links, tunnels and the dashboard are not part of the run, and neither is model behaviour.