Monitored Devices
Configure six monitoring sources, shared connections and machine history.

NasDash supports six monitoring sources reachable from its server. Save a shared connection in Settings → Integrations. Single-machine sources offer automatic machine creation, enabled by default. For Beszel, Prometheus and Proxmox, test and select discovered machines. Manage machines in Settings → Widgets → Devices, selecting their connection and target; credentials stay with the connection.
Sources and connection fields
| Source | Usual port | Connection and target |
|---|---|---|
| Glances | 61208 | Web API address; optional Basic authentication |
| Netdata | 19999 | Agent address; optional Basic authentication |
| Beszel | 8090 | Hub address, email and password; select a system |
| Prometheus | 9090 | Prometheus server; optional Basic or bearer authentication; select a node_exporter instance |
| Proxmox VE | 8006 | Hostname without scheme, port or path; separate Port, Token ID and secret; select node, VM or LXC |
| Libre Hardware Monitor | 9001 | Host, optionally scheme-prefixed, without port or path; separate Port; no authentication |
Refresh and history
The default monitoring interval is ten seconds while a browser is active and sixty seconds when idle. Configure it in Settings → Tasks and backups. Machine views include recent and 24-hour metric history. Add fleet, single-machine or metric widgets to any page; deleting a widget keeps the machine and its connection.
For source setup, see Glances, Netdata, Beszel, Prometheus/node_exporter, Proxmox VE and Libre Hardware Monitor. See Troubleshooting for errors.
CPU and GPU temperatures are separate readings in device and metric widget settings. The existing Temperature selection keeps its CPU meaning. GPU temperature becomes selectable only when the monitoring source reports it; GPU detection alone cannot supply a missing sensor. Each reading keeps its own visibility, display style, colour and threshold. A failed collection marks the source unavailable while preserving the last reading as stale, rather than presenting it as a fresh online measurement.
Sensor availability depends on the provider and hardware. If a CPU sensor is missing, CPU temperature remains unavailable; NasDash does not substitute the hottest GPU, disk or unspecified sensor for it.
Measure names and layout
Measures are named by their kind only: CPU, RAM, GPU, iGPU, Disk and temperatures. A disk or GPU is named individually only when a machine has several; its full name stays visible in the tooltip. Integrated graphics (Intel UHD or Iris, AMD Radeon Graphics or 7xxM, Apple and Jetson) are shown as iGPU. When a machine has both an iGPU and a GPU, the single-machine widget settings list both under GPU so either can be hidden; the fleet widget shows the dedicated GPU.
Names sit on the left and values on the right. When they do not fit on one line, the values move to the next line, spread from edge to edge. A name too long for its line is shortened with an ellipsis rather than wrapped. In one-measure widgets (Storage, GPU…), the machine name stays visible next to a disk or GPU name.