Skip to main content
The CA Colombia API enforces rate limits using fixed-window counters to protect service availability. Each endpoint tracks requests per IP address within a defined time window, and if you exceed the allowed number of requests that window resets before you can continue. Limits vary by endpoint — public and user-facing endpoints use a higher default, while staff-only endpoints apply a stricter cap.

Default Limits

Most endpoints apply the following default rate limit: Staff-only endpoints (such as character approval and review workflows) apply a stricter limit:
Rate limits are applied per IP address. All requests originating from the same IP — regardless of which user account is authenticated — share the same quota for a given endpoint.

Rate Limit Exceeded Response

When you exceed the allowed number of requests within a window, the API returns an HTTP 429 response with the following body:
You should pause new requests until the current window has reset before retrying. The response does not currently include Retry-After headers, so you should implement your own backoff strategy (see Best Practices below).

Roles That Bypass Rate Limits

Accounts with any of the following roles are exempt from rate limiting and will never receive a 429 response:
If your application requires sustained high-volume access, contact the CA Colombia team about obtaining a BOT role for your integration account.

Best Practices

Implement exponential backoff when you receive a 429 response. Start with a short pause (for example, 1 second), then double the wait time on each successive retry up to a reasonable maximum (for example, 30 seconds). This prevents your client from hammering the API immediately after a window resets and triggering another rate limit.
  • Cache responses that are unlikely to change frequently (such as user profile data) to reduce the number of requests you make within each window.
  • Monitor 429 responses in your logging and alerting pipelines so you can identify when a part of your integration is approaching its limit before users are impacted.