Skip to main content

Error format

details can contain additional information. Log the status, error code, and X-Request-Id without recording the secret token.

Rate limits

The default budget is 60 requests per minute per token; individual tokens can have a custom limit. Authenticated responses include X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset. For 429, use Retry-After as the waiting instruction; do not create a tight retry loop based on the reset header. Reduce concurrency and group reads appropriately. Retry writes only after checking their outcome. Repeating invalid inputs or requests without the necessary permissions does not resolve them.