Crate ritlers 

A task-length aware rate limiter.

Most rate limiters count when a task starts. Because network requests have variable latency, this can still cause bursts to arrive at a downstream API simultaneously. Ritlers instead waits for each task to finish before its slot becomes available again, so the rate limit is respected regardless of how long individual requests take.

When tasks complete instantly the behaviour is identical to a standard token-bucket limiter.

Feature flags

Feature	Default	Description
blocking	yes	Synchronous blocking::RateLimiter that parks the calling thread
async	no	Async async_rt::RateLimiter backed by Tokio

Quick start

Blocking

use std::time::Duration;
use ritlers::blocking::RateLimiter;
use ritlers::TaskResult;

let mut limiter = RateLimiter::new(2, Duration::from_secs(1)).unwrap();

// Schedule a plain task
limiter.schedule_task(|| { /* perform API call */ });

// Schedule a task that retries on rate-limit errors
limiter.schedule_task_with_retry(|| {
    match do_api_call() {
        Ok(_)  => TaskResult::Done,
        Err(_) => TaskResult::TryAgain,
    }
});

Async

use std::time::Duration;
use ritlers::async_rt::RateLimiter;
use ritlers::TaskResult;

let limiter = RateLimiter::new(2, Duration::from_secs(1)).unwrap();

// Schedule a plain task
limiter.schedule_task(async { /* perform API call */ }).await;

// Schedule a task that retries on rate-limit errors
limiter.schedule_task_with_retry(|| async {
    match do_api_call().await {
        Ok(_)  => TaskResult::Done,
        Err(_) => TaskResult::TryAgain,
    }
}).await;

Modules

async_rt

blocking

Structs

ZeroAmountError

Error returned by blocking::RateLimiter::new and async_rt::RateLimiter::new when amount is zero.

Enums

TaskResult

The outcome of a retryable task.