RateLimiter
An interface that provides rate-limiting functionality for executing operations.
Implementations of this interface decorate a suspend function call, controlling the rate at which it can be executed. If the call rate exceeds the configured limit, the execute method will either suspend until permission is granted or throw a RateLimitExceededException.
Semantics (token-bucket): The default implementation (DefaultRateLimiter) uses a token-bucket strategy. Up to RateLimiterConfig.maxCalls tokens are available; each execute consumes one. Tokens refill by adding maxCalls tokens at the start of each RateLimiterConfig.period (fixed window refill). For example, with maxCalls=10 and period=1s, at most 10 calls per second are allowed; after 1s from the last refill, the bucket is refilled with 10 tokens. This is a per-policy limit (single bucket). For per-key or per-tenant limits, a custom implementation or wrapper would be needed.