upstash-ratelimit-rs

Crates.ioupstash-ratelimit-rs
lib.rsupstash-ratelimit-rs
version1.0.2
sourcesrc
created_at2024-03-22 17:17:28.573643
updated_at2024-03-30 06:25:38.507363
descriptionAn unofficial Upstash rate limiting SDK in Rust
homepagehttps://github.com/sourabpramanik/upstash-ratelimit-rs
repositoryhttps://github.com/sourabpramanik/upstash-ratelimit-rs
max_upload_size
id1182817
size91,066
Sourab Pramanik (sourabpramanik)

documentation

README

Unofficial Upstash rate limit SDK for Rust

A rate-limiting SDK built for the Rust ecosystem that uses in-memory data storage.

Inspiration

This rate limit SDK is inspired by the official TypeScript rate limit SDK created by Upstash team.

Setup

  1. To setup the ratelimiter, first create a client instance of Redis that can store the request counts for a given set window:
let connection_str = std::env::var("UPSTASH_REDIS_URL").unwrap_or_else(|_| panic!("Expecting UPSTASH_REDIS_URL to be set"));
let Ok(redis) = redis::Client::open(connection_str) else {
    panic!("Failed to connect")
};
  1. Create a client instance of the RatelimitConfiguration using the Redis client:

  2. Use the client configuration to create a new instance of any one of the three rate-limiting algorithms:

For example: Using the fixed window algorithm to limit 10 requests in 30 seconds of the window.

let client = RatelimitConfiguration::new(redis, true, Some(String::from("my-custom-prefix")));
let ratelimit = FixedWindow::new(client, 10, "30s");

In the above client configuration, using the Ephemeral cache to avoid making Redis calls if the request is already blocked and adding a custom prefix string will override the default prefix string,

Use the ratelimit instance to call the limit function in any request calls to rate limit your requests:

let limit_response = state.ratelimit.limit("some-unique-identifier-like-ip", None).await;

Custom rate

By default every algorithm consumes one token per request, but if you want rate-limit the requests based on the payload size or any other factor, you can do so by providing the rate value to the limit function call:

let limit_response = state.ratelimit.limit("some-unique-identifier-like-ip", Some(10)).await;

This will consume 10 tokens in one request.

Examples

Check the examples directory

Roadmap

  • Single Region (may have latency issues)

    • Fixed window algorithm ✅
    • Sliding window algorithm ✅
    • Token bucket algorithm ✅
    • Cached fixed window algorithm 🛠️
    • Analytics 🛠️
    • Forced timeout 🛠️
    • Hard reset 🛠️
  • Multiple Region (no latency issues)

    • Fixed window algorithm 🛠️
    • Sliding window algorithm 🛠️
Commit count: 68

cargo fmt