caches

The caches clause defines one or more named in‑memory caches. Caches allow pipelines (actions, backends, scripts, etc.) to store and reuse values for a limited time, reducing repeated backend lookups and improving overall performance.

Each cache sub-block declares a logical key space where keys map to a collection (vector) of values. Entries can optionally expire automatically if a timeout is configured, or per insertion when values are written.

Only the metadata (cache names and timeouts) is configured here. The actual insertion, lookup, and invalidation operations occur at runtime via pipeline actions (e.g. actions that add or read cached values inside AAA processing), Lua scripts, or management APIs.

Syntax

caches {
    cache "NAME" {
        # Optional default expiration
        timeout 60s;
    }

    cache "SESSIONS" {
        timeout 30m;  # 30 minutes
    }
}

Cache block

A cache block supports:

If timeout is omitted, keys inserted without an explicit timeout will persist until explicitly removed or the process restarts.

See these cache usage examples:

• Lua cache API documentation - Lua cache API usage, including rate limiting.
• @init pipeline - Use the cache action to pre-populate cache entries when configuration loads or reloads.
• cache action - Store values in a configured cache.
• totp action - Read values from cache and update cache entries for TOTP replay protection.

Example

caches {
    cache "USER_ATTRIBUTES" {
        timeout 5m;  # 5 minutes
    }

    cache "SESSION_DATA" {
        timeout 30m;  # 30 minutes
    }

    cache "RATE_LIMIT" {
        timeout 1h;  # 1 hour
    }
}