Metrics

Read live resource usage for a sandbox — CPU, memory, process count, network, and uptime.

PythonJavaScript

python

m = sbx.get_metrics()
print(m["cpu_pct"], m["mem_used_mb"], m["uptime_s"])

javascript

const m = await sbx.getMetrics();
console.log(m.cpu_pct, m.mem_used_mb, m.uptime_s);

Fields

Field	Meaning
`status`	`running` or `paused`
`cpu_pct`	CPU usage percent (point-in-time)
`cpu_cores`	cores visible to the sandbox
`mem_used_mb`	memory in use (page cache excluded)
`mem_limit_mb`	memory limit
`mem_pct`	memory used as a percent of the limit
`pids`	running process count
`net_rx_mb` / `net_tx_mb`	network received / transmitted
`uptime_s`	seconds since the sandbox was created

A paused sandbox reports zero compute (its container is stopped).

Poll get_metrics on a timer to drive a live usage view — the AI2IN dashboard does exactly this in the sandbox workspace.