npm/spife-ratelimit

Name: spife-ratelimit

Owner: npm

Description: :alarm_clock: :no_entry_sign: ratelimit spife apps with token buckets!

Created: 2017-08-25 23:15:49.0

Updated: 2018-04-24 01:21:50.0

Pushed: 2018-03-12 16:41:05.0

Homepage:

Size: 59

Language: JavaScript

GitHub Committers

User	Most Recent Commit	# Commits

Other Committers

User	Email	Most Recent Commit	# Commits

README

spife-ratelimiter

Middleware that adds Token Bucket-style ratelimiting to Spife.

ettings.js:
le.exports.MIDDLEWARE = {
@npm/spife-ratelimiting', {
redisUrl: 'redis://localhost:6379', // let it connect to redis
redisClient,                        // (or give it a client)

costHeader: 'npm-cost',             // the "cost" header to use
                                    // for determining how many
                                    // tokens a request took

clock: Date.now,                    // if you would like time to
                                    // move to the beat of a
                                    // different drum

getBucket (spifeRequest) {          // given a request, return
                                    // an object describing its
                                    // token bucket
  return {
    capacity: 100,                  // how many tokens does
                                    // the bucket hold?

    interval: 60,                   // in minutes, how long
                                    // does the bucket take to
                                    // fill up?

    key: String                     // what should the bucket be
                                    // named?
  }
}



iews.js:
t reply = require('@npm/spife/reply')
t cost = require('@npm/spife-ratelimiter/decorator')

le.exports = {
 a view can declare how much it costs up-front:
dex: cost(index, 5),
her


tion index (req, context) {
turn 'I cost five tokens!'


tion other (req, context) {
turn reply.header(
'I cost a hundred tokens sometimes',
'npm-cost',
Math.random() > 0.5 ? 100 : 0

This is implemented in redis by keeping two keys per bucket: one that tracks the current number of tokens, and one that tracks the last update time. Both keys are expired at interval, and updated atomically using a Lua script.