ratecontrol
v1.1.0
Published
A simple rate controller for connect-style servers.
Downloads
4
Maintainers
andyburkeReadme
ratecontrol
A simple rate controller for connect-style servers.
Code Example
const RateControl = require( 'ratecontrol' );
// use a global rate control
const globalRateControl = RateControl( {
rate: '10/s'
} );
server.get( '/rate_limited', globalRateControl, onGetLimited );
// or use a specific one for a specific route
server.get( '/specific_rate_limited', RateControl( {
rate: '1/s'
} ), onGetSpecificLimited );Motivation
A simple rate controller is nice to have.
Installation
npm install --save ratecontrolOptions Reference
rate
A human-readable rate setting, egs:
server.get( '/10_per_second', RateControl( {
rate: '10/s'
} ), onRequest );
server.get( '/1_per_second', RateControl( {
rate: '1/second'
} ), onRequest );
server.get( '/1000_per_day', RateControl( {
rate: '1000/day'
} ), onRequest );trustedHeaders
An array of headers to trust for determining the source IP used for rate limiting. Default:
[
'x-client-ip',
'x-forwarded-for',
'x-real-ip',
'x-cluster-client-ip',
'x-forwarded',
'forwarded-for',
'fowarded'
]Be wary of this setting. It is very liberal by default, which could allow a malicious actor to spoof a changing IP and get around the rate control. This liberal setup assumes that you have set up your app behind a sane proxy. If you'd like to lock this down, it's an easy configuration change:
server.get( '/locked_down', RateControl( {
rate: '10/s',
trustedHeaders: []
} ), onLockedDown );getId
You can override the method which determines the id of the client:
server.get( '/override_getid', RateControl( {
rate: '10/s',
getId: ( request, callback ) => {
callback( null, request.user.id );
}
} ), onOverrideGetId );By default, ratecontrol uses the IP address for rate limiting. First, the list of trusted headers is checked, then request is examined to extract an IP. If no id can be found, rate limiting is not applied.
cache
Allows you to override the default cache. Must be an object with asynchronous get/set methods that return a rate limiter with a consume() method. Eg:
// let's use redis so we can do rate limiting
server.get( '/overridden_cache', RateControl( {
rate: '10/s',
cache: {
get: ( id, callback ) => {
redis.get( id, ( error, info ) => {
if ( error ) {
callback( error );
return;
}
return TokenPipeline( info );
} );
},
set: ( id, limiter, callback ) => {
redis.set( id, limiter.toJSON(), callback );
}
}
} ), onOverriddenCache );By default ratecontrol uses a small LRU in-memory cache.
errorBody
You can override the body of the response when the rate limit is exceeded:
server.get( '/custom_error_body', RateControl( {
rate: '10/s',
errorBody: '<html><body><blink>RATE LIMIT EXCEEDED</blink></body></html>'
} ), onCustomErrorBody );By default, the response is a json object:
{
"error": "rate limit exceeded",
"message": "Too many requests."
}onRateLimitExceeded
You can override the method called when the rate limit is exceeded:
server.get( '/custom_rate_limit_method', RateControl( {
rate: '10/s',
onRateLimitExceeded: ( request, response, next ) => {
mySecretLog.log( `rate limit exceeded for customer: ${request.customer}` );
next(); // don't actually produce a user-visible error
}
} ), onCustomRateLimitMethod );Tests
npm run testContributing
Contributions are encouraged and appreciated. To make the process as quick and painless as possible for everyone involved, here's a checklist that will make a pull request easily accepted:
- Implement your new feature or bugfix
- Add or update tests to ensure coverage
- Ensure your code passes jshint according to the .jshintrc
- Ensure your code is formatted according to the .jsbeautifyrc
- Submit
License
MIT