hirocli metrics provides terminal access to the workspace server’s resource monitoring. It communicates with the running server over HTTP, so the server must be running with metrics enabled.
Metrics collection is disabled by default. Enable it persistently with
hirocli setup --metrics or for a single run with hirocli start --metrics. See Setup and uninstall and Server ops.status
- whether the collector is enabled
- collection interval in seconds
- ring buffer capacity and current fill count
- how long the collector has been running
Options
| Option | Description |
|---|---|
--workspace, -W | Target workspace name. |
Example
snapshot
- Server process — CPU percent, RSS memory, virtual memory, thread count
- Channel plugins — per-plugin CPU, RSS, thread count, alive status
- Disk — usage percent, read rate, write rate
- Network — bytes sent/received per second, packet rates
- System-wide — total CPU percent, memory used/total
Options
| Option | Description |
|---|---|
--workspace, -W | Target workspace name. |
Example
live
GET /metrics and refreshes the display in place. Press Ctrl+C to stop.
The dashboard is rendered with rich and uses the same table layout as snapshot, refreshing at the specified interval.
Options
| Option | Description |
|---|---|
--workspace, -W | Target workspace name. |
--interval | Refresh interval in seconds. Defaults to 2. |
Example
Enabling metrics
Metrics collection is controlled at two levels:Persistent (written to config.json)
Ephemeral (for a single server run)
--metrics flag on start and restart behaves the same as --admin: it overrides the config value for that server run without persisting the change.
Runtime configure
While the server is running, you can toggle metrics or change the interval without restarting:For architecture details — how the collector is implemented, what each metric measures, and the REST API contract — see Metrics monitoring.
