linagirl / leaky-bucket Goto Github PK

I'm trying to use this package in a client-side build. I'm unable to do so because .mjs files are not supported (out of the box, at least). Renaming the files to .js works, and the import can be found, but the module also uses logd, which is node.js only as it uses fs.

I made PR that uses rollup to compile both .js and .ejs and replaces logd with debug. Let me know what you think!

There's a typo in your dockets (a "bucket" is misspelled)

hi
i need use leaky bucket in express app for api and i do not know how do it
please help me
thanks

On a mac.

bash-3.2$ node -v
v8.9.4
bash-3.2$ npm -v
5.6.0
bash-3.2$ npm install leaky-bucket
npm WARN [email protected] No repository field.
npm WARN [email protected] No license field.

• [email protected]
  added 71 packages in 3.584s
  bash-3.2$ node

var LeakyBucket = require('leaky-bucket');
TypeError: path must be a string or Buffer
at Object.fs.statSync (fs.js:948:11)
at classConstructor.init (/Users/davidh/WebstormProjects/webClientSimulator/node_modules/ee-log/lib/Logger.js:59:15)
at new classConstructor (/Users/davidh/WebstormProjects/webClientSimulator/node_modules/ee-class/src/class.js:162:47)
at /Users/davidh/WebstormProjects/webClientSimulator/node_modules/ee-log/lib/Logger.js:52:12

If the capacity is 100 and the interval is 60 seconds and a request has a cost of 1, every 60 seconds 1000 request may be processed. If the request cost is 4, jsut 25 requests may be processed every 60 seconds.

Did taht first sentence mean to say every 60 seconds 100 requests may be processed?
If the README is correct, why is it only 25 if the cost becomes 4?

Hi, I have such config:

new LeakyBucket({
    capacity: 1000,
    interval: 60,
    timeout: 300,
})

Version of the leaky-bucket package I use: v3.0.4

The API server returns wasted cost on each request. Each request has 50 points cost (await bucket.throttle(50)). For some reason first wave of my requests break the leaky bucket, remote API server returns 1500 wasted points and I'm getting ban. If the service doesn't ban me, then the bucket starts to work properly and the remote API returns wasted points stays below capacity all the rest time.

I'm not sure how to write a test for the case. But I found a workaround, right after LeakyBucket creation I pay full capacity points and it start to work as it is expected:

const capacity = 1000;
const bucket =  new LeakyBucket({
    capacity,
    interval: 60,
    timeout: 300,
});

bucket.pay(capacity); // if you remove this line the remote api service will ban you

(async () => {
    for (let i = 0; i < 5000; ++i) {
        await bucket.throttle(50);
        const response = await fetch(...)
    }
})();

P.S. sorry for my English