Rate Limiter Node JS

A simple way to limit how often a function is executed.

Currently works with node.js v0.10.1+ (and probably lower).

Examples

If you want to limit all requests:

var limit = require("rate-limiter-node");
var request = limit(require("request")).to(10).per(1000);

userIds.forEach(function(userId) {
	var url = "http://anywebsite/users/" + userId;
	request(url, function(err, res, body) {
		/* ... Yay! Not a too-many-request-per-second error! ... */
	});
})

Or if you'd like to be a bit more fine grain and/or explicit:

var limit = require("rate-limiter-node");
var callApi = limit(function(userId, callback) {
	var url = "http://easily-overwhelmed-api.com/users/" + userId;
	request(url, callback);
}).to(10).per(1000);

userIds.forEach(function(userId) {
	callApi(userId, function(err, res, body) {
		/* ... Yay! Not a too-many-request-per-second error! ... */
	});
})

API

Basic usage: var limited = limit(fn);

• limited(args...): Enqueues a set of arguments that fn will be executed with. Returns an EventEmitter that will emit a limiter-exec event when fn is executed with args. Also, in the case that fn returns an EventEmitter, the events emmitted by the returned EventEmitter will be relayed to the EventEmitter returned by limited. In other words, stuff like this works:

  var limit = require("rate-limiter-node");
  var request = require("request");
  limit(request)("http://google.com").on("data", function(chunk) { /*** code ***/ });

• limited.to(count=1): The number of times fn will be execute within the time specified by .per(time).

• limited.per(time=Infinity): The period of time in which fn will be executed .to(count) number of times.

• limited.evenly([toggle=true]): Off by default. When true, fn will be executed evenly through the time period specified by .per(time). For example, if set to true and .to(10) and .per(1000), then fn will be executed every 100ms.

• limited.withFuzz([percent=0.1]): Set to 0 by default. Adds a random factor to the delay time. For example if set to 0.1 and .to(10) and .per(1000), then fn will be executed between every 100ms to 110ms.

All methods attached to limited and can be chained.

Edge cases

Basic usage: limit(fn).to(count).per(interval)

The to method or per can be called any number of times including zero. The effective default count for to is 1 and the effective default interval for per is Infinity. Which means if neither the to method or per method are called, then limit will only execute fn once ever:

var init = limit(function() { console.log("Runs only once"); });
for(var i = 0; i < 3; i++) { init(); }

/*** Console output: ***/
// $ Runs only once
// ... Nothing else is ever printed ...

Similarly, if per is never called, fn is executed at most count number of times:

var dos = limit(function() { console.log("twice only"); }).to(2);
for(var i = 0; i < 3; i++) { dos(); }

/*** Console output: ***/
// $ twice only
// $ twice only
// ... Nothing else is ever printed ...

On the other hand, if to is never called, fn is only executed every interval milliseconds:

var tick = limit(function() { console.log("Once a second..."); }).per(1000);
for(var i = 0; i < 3; i++) { tick(); }

/*** Console output: ***/
// $ Once a second
// ... pause for 1 second ...
// $ Once a second
// ... pause for 1 second ...
// $ Once a second

Finally, as mentioned earlier, to and per can be called any number of times. Each time either to or per is called, the previous count or interval value is overriden:

var strange = limit(function() {
	console.log("Every once in a while");
}).to(1).per(1000).to(2).per(2000);
for(var i = 0; i < 3; i++) { strange(); }

Getting rate-limiter-node

The easiest way to get simple-rate-limiter is with npm:

npm install simple-rate-limiter

Alternatively you can clone this git repository:

git clone https://github.com/omair445/request-limiter-node-js.git

Developed by

• Omair Afzal --> Otif Solution

License

This project is released under The MIT License.