Management API Limits
To ensure stability, the Relay Management API enforces the following rate limits across all workspaces based on IP address and authentication token:
- Authentication endpoints: 20 requests per 15 minutes.
- Standard endpoints (e.g. creating/listing APIs): 100 requests per 1 minute.
- Sensitive endpoints (e.g. revoking API keys): 10 requests per 1 minute.
Gateway Limits
For your actual API targets, rate limiting is opt-in. When you create an API via the Create API Route, you can specify rate_limit_enabled: true, along with a rate_limit_max and rate_limit_window. These limits are evaluated per unique API Key making requests through the gateway.
Handling Rate Limits
When a rate limit is exceeded, Relay returns an HTTP 429 Too Many Requests response. The response includes standard rate limit headers to help you gracefully back off:
X-RateLimit-LimitThe maximum number of requests allowed in the current window.
X-RateLimit-RemainingThe number of requests remaining in the current window.
X-RateLimit-ResetThe Unix timestamp when the current window expires and limits are reset.