Skip to main content
Version: 1.0.x

Custom Quota Strategy

The Custom Quota strategy provides granular control over API traffic, enabling limits based on tokens and requests within a fixed time window.

Supports configurations for:

  • Maximum requests allowed
  • Flexible time windows for rate limiting (e.g., minute, hour).
 
/etc/lunar-proxy/quotas/{fileName}.yaml
quotas:
  - id: FixedWindowCustomCounter # Unique identifier for the quota
    filter:
      url: api.openai.com/* # URL pattern to apply the quota
    strategy:
      fixed_window_custom_counter:
        max: 40000  # Maximum requests allowed within the window
        interval: 1
        interval_unit: minute # Time window duration in minutes
        counter_value_path: | # JSONPath used to define where to collect token count
          $.request.headers["x-lunar-used-tokens"]