Skip to main content

Telemetry

Version 0.1.8.

The extensions marketplace lists a separate Resource Telemetry Collector entry, badged AGENT PLUGIN — that's the device-side collector plugin this view installs and configures on a device, not this desktop extension.

What it does​

Telemetry charts the metrics a device reports — CPU, memory, disk, network, and whatever else its collector plugin gathers — live, either for one device or several overlaid together. It reads from whichever source is configured: the organization's default storage device over the mesh, a cloud telemetry API, or mesh-first-with-cloud-fallback (the default).

When you'd use it​

Check a device's health at a glance, compare resource use across several devices side by side, change which metrics a device collects, or pull a CSV of a device's recent metric history.

Enable or install​

Like other extensions, it can be enabled, disabled, or uninstalled from the Extensions view. It contributes the Telemetry: Open command and a sidebar entry titled Telemetry.

Key screens​

  • Toolbar — device and online counts, a "Live · every 30s" badge once devices are selected, time-range buttons (1h / 6h / 24h / 7d), and Refresh.
  • Device rail — one row per enrolled device, multi-selectable (at least one stays selected); a device reporting nothing shows a dimmed "no data" badge.
  • Single-device view — the device's name and a health pill; a meta strip (host, uptime, last seen, updated); a collapsible Metrics selector and a Flight recorder export control under the header; then its metrics as cards grouped into Compute / Network / Storage / Other sections with filter chips, each expandable into a larger view.
  • Multi-device view (2 or more selected) — the same family-grouped sections, but each card overlays one line per selected device, with a color-coded legend.
  • Metrics selector — collapsed by default; one row per metric the device's collector plugin can gather (its emitted name, its underlying key, and roughly what it costs — about 0.5 MB/day per enabled metric), with Apply. If the collector plugin is missing, disabled, or crashed, this area explains which and — where a matching package is staged on your desktop — offers Install or Reinstall.
  • Flight recorder — exports a window (1, 5, or 15 minutes) of the device's on-device metrics buffer as a CSV, via Export CSV and a folder picker, with progress and a note that the export contains collector-plugin metrics only, not the daemon's own host samples.

Common tasks​

  • Check one device — select it in the rail; expand a card for a larger chart.
  • Compare devices — select more than one; the same sections switch to overlaid, color-coded charts.
  • Change what's collected — open Metrics under the device's header, tick or untick, Apply.
  • Install the collector plugin — if a device reports it's missing, use the Install button (only available when a matching package is staged locally).
  • Export recent history — open Flight recorder, pick a window, Export CSV, choose where to save.

Permissions requested​

commands.register, notifications.show, devices.read, authentication.session, daemon.files.

Known limits​

  • Nothing in the stack can read a collector plugin's configuration back, so an applied metric selection is only ever the desktop's own record of what it last sent — it can drift from what the device is actually doing.
  • Installing or reinstalling the collector plugin only works when a matching package happens to be staged on this desktop; otherwise the device is reported as unavailable, with no install option offered.
  • Exports cover collector-plugin metrics only; the daemon's own host samples (CPU, memory, disk, plugin count) travel a separate path and never appear in one.

Written from source; not verified against a live backend.

Fleet Manager — installs, enables, and disables the collector plugin this view depends on.