oz0 tenant cache
With auto, the gateway manages the prompt cache in three ways:
For: Operators and developers looking up a command.
With auto, the gateway manages the prompt cache in three ways:
- It marks the stable start of each request (the system prompt and the history before the platform's status block) for providers that cache only what a request marks, and only where a request has no marks of its own.
- It asks for the provider's longer cache lifetime when a call offers a tool that waits for a person and the longer lifetime's higher write price is likely to pay.
- When a person decides an approval after the cache lapsed, it re-sends the job's last request once as the smallest request the provider allows, metered on the job under its budget.
With off, calls go to the provider exactly as agents wrote them. Either way every call's misses are named (usage cache --causes).
Turning it off loosens no control, so it takes no reason; admins of the tenant may change it, and each change is audited.
Usage
oz0 tenant cache <tenant> auto|off [flags]
Flags
| Flag | Type | Default | Description |
|---|---|---|---|
--json | bool | print JSON instead of a table |
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.