Design a rate limiter for a microservices API.

A robust rate limiter for a microservices API should consider various identification methods beyond just IP addresses, such as JWT tokens, to accurately track and limit requests per user or service. A common approach involves using a data structure like a deque (double-ended queue) to store timestamps of recent requests associated with a specific identifier. By maintaining a count of requests within a defined time window (e.g., the size of the deque or a fixed duration), we can enforce rate limits. When a new request arrives, we add its timestamp to the deque and remove any timestamps that fall outside the window. If the number of timestamps in the deque exceeds the predefined threshold, the request is rejected.