charged machine in a dark hangar
telemetry polled from the machine every 5 minutes · auto-refresh 60s

LIVE STATUS

Snapshot · ·

GPU Temp
°C
GPU Power
W
GPU Use
%
GPU Mem Alloc
%
RAM Used
GiB
Load

GPU telemetry history (rolling 24h)

Services & endpoints

Benchmarks (live from the bench logs)

modelbackendtestt/s±

Model inventory on /srv/llm/models

Publisher: ashborn poller → Cloudflare Pages (no agents run on the machine; SSH read-only over the tailnet)