Real-time system metrics from a simulated 32-core / 512 GB / 1 GPU workstation.
How the test data is generated#
This public demo is a browser-only test page. Its panel configuration uses
api: "sample-stream", so the browser does not contact a backend service.
Every second, sample-stream.js produces the same data shape as the real
/metrics and /profile endpoints and passes it to the normal dashboard
renderer.
The values describe a simulated 32-core CPU, 512 GB of memory, one 24 GB GPU, and a 1 TB disk. CPU, memory, GPU, and temperature values vary smoothly from time-based signals. CPU busy-core selection uses a fixed seed, and the whole scenario repeats every 60 seconds. Storage capacity is fixed at 42% used. The deterministic loop makes the page safe to host as static files and gives the renderer realistic changing input without exposing a real machine.
Simulated Workstation
sample-streamDeploying with real data#
GitHub Pages can only host the static dashboard; it cannot run the
workstation-probe backend service. For a real deployment, build and run the
workstation-probe binary on each machine being observed, configure its
listening address and port, and serve it through HTTPS. For example:
make build
sudo make install
sudo systemctl enable --now workstation-probeThen change the panel config’s api value from "sample-stream" to the
workstation-probe service’s base URL:
name: "Production Workstation"
api: "https://workstation.example.com"
refresh: "5s"
modules:
- cpu
- memory
- gpu
- storageThe browser must be able to reach that URL. If the dashboard and
workstation-probe use different origins, enable CORS for the dashboard
origin; putting both behind the same reverse proxy avoids CORS. Protect the
service with the access controls required by the environment rather than
exposing an unauthenticated metrics endpoint to the public internet. Keep
sample-stream for demos and use a separate, environment-specific panel
config for production.