Vast Monitor — Fleet instrument Live reading

Hosts online

12/ 12

  • 9 online
  • 2 rented
  • 1 offline

Your GPU fleet, read like an instrument.

Vast Monitor puts every hosted machine — thermals, load, power, earnings — on one calibrated screen you can read from a phone.

What this instrument reads

One screen for the signals that decide whether a rented GPU keeps earning.

Group Signals
Compute CPU utilization · RAM usage · GPU utilization · GPU memory
Thermal + power GPU temperature · GPU power draw · System temperature · Trend-aware averages
Capacity + transit Disk utilization · Disk capacity · Network status · Fleet state changes

Method

Connect, stream, read. Three notes to a working instrument.

  1. Connect the account

    Link your Vast.ai machine/account so every hosted GPU is discoverable as a registered lane.

  2. Stream the telemetry

    The backend ingests time-series metrics through VictoriaMetrics and Grafana Alloy, live.

  3. Read it anywhere

    Glance at fleet health and earnings from the web or the Flutter app, wherever you are.

The live reading

Every host is a numbered lane, kept in line by its own state.

Fleet lanes 12 machines · synced
  • gpu-host-01 GPU 86% · 67°C · 248W
  • gpu-host-02 Rented · RAM 72% · Disk 61%
  • gpu-host-11 Offline · alert pending see alert on lane

Illustrative sample of the fleet view.

Next revision

Be the first to read your fleet from your pocket.

Joining the waitlist gets you a note when Vast Monitor ships to the Play Store.

Get notified at launch