A sandbox's current CPU, memory and disk use and its recent history
The control plane samples every running sandbox at a fixed interval
(interval_seconds) and keeps the samples of the last hour in memory;
they start over when the control plane restarts. current is
present while the sandbox runs and has a recent sample. A stopped
sandbox keeps the history of its last run until it ages out. Answers
503 when the deployment does not sample resources.
/v1/sandboxes/{sandbox}/statsAuthorizationBearer token · headerrequiredOrg-scoped API key, pols_....
sandboxstringrequiredSandbox ID (sbx_...) or name.
The sandbox's resource use.
sandbox_idstringrequiredinterval_secondsintegerrequiredHow often running sandboxes are sampled.
currentResourceStatsOne sample of a running sandbox's resource use, as stats on a
sandbox (present while it runs and has a recent sample) and in
GET /v1/sandboxes/{sandbox}/stats.
Show propertiesHide properties
sampled_atstring<date-time>requiredcpu_percentnumber<double>Share of the sandbox's vCPUs that were busy, averaged since the previous sample: 100 means all of them. Absent in the first sample after the sandbox starts.
cpu_coresnumber<double>The same as a number of busy vCPUs, for example 1.5. Absent with cpu_percent.
memory_used_bytesinteger<int64>requiredMemory in use inside the VM, as its guest agent reports it.
memory_total_bytesinteger<int64>requiredMemory the guest sees (slightly less than the size's RAM, which the guest kernel reserves part of).
disk_used_bytesinteger<int64>requiredSpace the root disk volume takes in the host's storage pool, as the pool reports it. On a copy-on-write clone (a sandbox created from a template, or a fork) this can leave out blocks it still shares with its origin. 0 when the host does not report it.
disk_total_bytesinteger<int64>requiredSize of the root disk.
historyResourceStats[]requiredThe samples of the last hour, oldest first, including current.
Show propertiesHide properties
ResourceStatssampled_atstring<date-time>requiredcpu_percentnumber<double>Share of the sandbox's vCPUs that were busy, averaged since the previous sample: 100 means all of them. Absent in the first sample after the sandbox starts.
cpu_coresnumber<double>The same as a number of busy vCPUs, for example 1.5. Absent with cpu_percent.
memory_used_bytesinteger<int64>requiredMemory in use inside the VM, as its guest agent reports it.
memory_total_bytesinteger<int64>requiredMemory the guest sees (slightly less than the size's RAM, which the guest kernel reserves part of).
disk_used_bytesinteger<int64>requiredSpace the root disk volume takes in the host's storage pool, as the pool reports it. On a copy-on-write clone (a sandbox created from a template, or a fork) this can leave out blocks it still shares with its origin. 0 when the host does not report it.
disk_total_bytesinteger<int64>requiredSize of the root disk.
Rate limited (rate_limited): too many requests or failed
authentications from this address, too many requests or
lifecycle calls for this org, or too many of its exec, file,
computer and CDP calls in progress at once. Retry after
Retry-After seconds.
errorobjectrequiredShow propertiesHide properties
codestringrequiredStable machine-readable code: bad_request (400),
unauthorized (401), forbidden (403), quota_exceeded
(403), not_found (404), conflict (409),
rate_limited (429, see Retry-After), internal (500),
runtime_error (502, the sandbox host failed),
unavailable (503, the feature is not configured on this
deployment), timeout (504, or 408 when a request body
stalls).
messagestringrequiredError.
errorobjectrequiredShow propertiesHide properties
codestringrequiredStable machine-readable code: bad_request (400),
unauthorized (401), forbidden (403), quota_exceeded
(403), not_found (404), conflict (409),
rate_limited (429, see Retry-After), internal (500),
runtime_error (502, the sandbox host failed),
unavailable (503, the feature is not configured on this
deployment), timeout (504, or 408 when a request body
stalls).
messagestringrequired