oz0 tenant sandbox-limits
For: Operators and developers looking up a command.
Every sandbox on a Linux node with sandbox tier 1 runs under memory, process and CPU limits, held by a cgroup the node's service delegates to it or by a systemd scope: a harness session gets 4096 MiB, 1024 processes and 2 CPUs, a plugin's MCP server 1024 MiB, 256 and 1, a command of its own (an ACP terminal, a test run) 2048 MiB, 512 and 2.
An agent asks for others with limits.memory_mb, limits.pids and limits.cpus in its definition. The ceiling set here is the most any one sandbox of the tenant gets, whatever its agent asks, and lowers a default above it. A sandbox that reaches its memory limit is stopped, all of it; one that reaches its process limit cannot start more.
The flags given replace the whole ceiling (a field not given has none); with no flags the ceiling is shown, and --clear removes it.
Usage
oz0 tenant sandbox-limits <tenant> [flags]
Flags
| Flag | Type | Default | Description |
|---|---|---|---|
--clear | bool | remove the ceiling | |
--cpus | float | the most CPU time, in cores (0: no ceiling) | |
--json | bool | print JSON instead of a table | |
--memory-mb | int | the most memory, in MiB (0: no ceiling) | |
--pids | int | the most processes and threads (0: no ceiling) |
Inherited flags
| Flag | Type | Default | Description |
|---|---|---|---|
--context | string | $OZ0_CONTEXT or <config dir>/orchestrator-zero/context.json | CLI context file written by server init |
--log-level | string | info | log level: debug, info, warn or error |
See all oz0 commands and the global flags.