Rate-limit Scheduler

This library helps you respect the rate-limits that services apply to their APIs.

Consuming APIs of 3rd party services often requires to deal with rate limits (such as that of Github). Properly managing these constraints is not an obvious task and if done improperly, your service will most likely be forbidden access to said service.

This library provides your code with a scheduler that will hold requests in respect to the rate-limits that the services you are consuming apply.

Installation

$ npm install --dev rate-limit-scheduler

Usage

In order to use the scheduler, you will need to write 2 functions:

• an action: typically, a function performing the API call
• an updater: a function that uses the return of the action to update the rate limits (typically, using headers or metadata)

You will find a demo of the scheduler with the Github API in the file demos/github.js.

Generally speaking, you will do:

let RateLimitScheduler = require('rate-limit-scheduler')
let scheduler = new RateLimitScheduler()
 
let result = await scheduler.schedule(actionFn, rateUpdaterFn)

The action function actionFn

This function is your actual call to your rate-limited service. Async functions are supported.

async function actionFn() {
  return await fetch('https://api.a-service.com/me?token=ABCDEFGH')
}

The update function rateUpdaterFn

This function will update the limits that the scheduler will adhere to.

function rateUpdaterFn(res) {
  return {
    limit: res.headers['x-ratelimit-limit'],
    reset: res.headers['x-ratelimit-reset'] * 1000,
    remaining: res.headers['x-ratelimit-remaining']
  }
}

The following properties are expected:

• limit: (optional) The maximum number of requests per timeframe.
• reset: An UNIX Epoch time indicating when the counter of allowed requests will be reset to limit.
• remaining: The number of remaining requests before being refused to access the API.

License

Licensed under the MIT License.

See LICENSE.md

© 2019 PeopleDoc SAS and the contributors