DEV Community

Discussion on: Limit concurrent asynchronous calls

Collapse
 
smishr4 profile image
Shubham Mishra

I implemented this whole thing with a working code, let me know if I am doing something incorrectly.

Note: I am using setTimeout for async calls so promise failure will not happen.

const ASYNC_LIMIT = 2;

function scheduler(cb, id, delay) {
    return new Promise((resolve, reject) => {
        setTimeout(() => {
            cb(id);
            resolve();
        }, delay);
    });
}

const asyncLimit = (fn, n) => {
    let promiseArray = [];

    return async function(...args) {
        if (promiseArray.length >= n) {
            await Promise.race(promiseArray);
        }

        let p = fn.call(this, ...args);
        promiseArray.push(p);

        p.then(() => {
            promiseArray = promiseArray.filter(pending => p !== pending);
        });

        return p;
    };
};

let cb = id => {
    console.log(id + " task completed", Date.now() % 10000);
};

let modifiedScheduler = asyncLimit(scheduler, ASYNC_LIMIT);

modifiedScheduler(cb, 1, 5000);
modifiedScheduler(cb, 2, 2000);
modifiedScheduler(cb, 3, 1500);
modifiedScheduler(cb, 4, 3000);
modifiedScheduler(cb, 5, 4000);
modifiedScheduler(cb, 6, 1000);
modifiedScheduler(cb, 7, 2500);

Collapse
 
tconrado profile image
tconrado

using the corrections from @kusha and the data and functino of your code it did work as expected!
hurray!

exceptional implementation to deal with REST API

Collapse
 
smishr4 profile image
Shubham Mishra

Can you post your working snippet here?

Thread Thread
 
tconrado profile image
tconrado • Edited

hey I'm terrible with markdown, but the example bellow is in a lib, basically it does limit the number of active connections to a http API rest service to the a defined number (8 in this case), so a the http request is resolved, it start another connection keeping always 8 active connections to the API server; it will completely hide the connection/handshake delay while guarantee no 429 error (too many requests); from my experience fastest safe approach as you can know the maximum number of calls per second of the API service


// this is the asyncLimit adjusted 
const asyncLimit = (fn, n) => {
  const pendingPromises = new Set();
  return async function(...args) {
    while (pendingPromises.size >= n) {
      await Promise.race(pendingPromises);
    }
    const p = fn.apply(this, args);
    const r = p.catch(() => {});
    pendingPromises.add(r);
    await r;
    pendingPromises.delete(r);
    return p;
  };
};

// native node.js https module to connect to shopify servers
const https = require('https')
exports.httpRequest = function(method, path, body = null) {
  const reqOpt = { 
    method: method,
    path: '/admin' + path,
    hostname: 'xxxxxxxxxxxxxxxxxxxx.myshopify.com', 
    headers: {
      "Content-Type": "application/json",
      "X-Shopify-Access-Token": "xxxxxxxxxxxxxxxxxxxx",
      'Cookie': '',
      "Cache-Control": "no-cache"
    }
  }
  if (body) reqOpt.headers['Content-Length'] = Buffer.byteLength(body);
  return new Promise((resolve, reject) => {

      const clientRequest = https.request(reqOpt, incomingMessage => {
          let response = {
              statusCode: incomingMessage.statusCode,
              headers: incomingMessage.headers,
              body: []
          };
          let chunks = ""
          incomingMessage.on('data', chunk => { chunks += chunk; });
          incomingMessage.on('end', () => {
              if (chunks) {
                  try {
                      response.body = JSON.parse(chunks);
                  } catch (error) {
                      reject(error)
                  }
              }
              resolve(response);
          });
      });
      clientRequest.on('error', error => { reject(error); });
      if (body) { clientRequest.write(body)  }  
      clientRequest.end();

  });
}


// the number 8 bellow can be changed to match the REST API service limits
// assume that this amount will call at once and will be replaced dynamically, hence
// if the service limit it 20 calls per second, be aware that 8 calls will hit the service at once
// using 40% of the maximum (avoid going higher)

exports.ratedhttpRequest = asyncLimit(exports.httpRequest, 8);

Collapse
 
tconrado profile image
tconrado

hey, tried your code, did not work...

so, assigned the same delay for all of those...like 2 seconds... I did expect to see it process 2, schedule + 2; so, the result would be observable every 2 seconds and everytime 2 task