Rate Limits

Per-tenant and per-principal token-bucket limits live at the gateway. When you exceed a bucket, the gateway returns:

HTTP/1.1 429 Too Many Requests
Retry-After: 5

{
  "error": {
    "code": "rate_limit.exceeded",
    "status": 429,
    "retriable": true,
    "retryAfter": 5
  }
}

Clients must honor Retry-After. The generated SDK (@ghasi/sdk-client) does this automatically with bounded exponential backoff + jitter.