如何在节点中管理大量HTTP请求

Question

Been looking around the net for an answer to this, but not found anything conclusive.

I have a node application that (potentially) needs to make a large number of HTTP GET requests.

Let's say http://foo.com/bar allows an 'id' query parameter, and I have a large number of IDs to process (~1k), ie
http://foo.com/bar?id=100
http://foo.com/bar?id=101
etc.

What libraries that folks have used might be best suited to this task?

I guess I'm looking for something between a queue and a connection pool:

1. The setup:
   

   • A large array of IDs exists to be processed (up to ~1k IDs)

2. The process:
   

   • Some kind of pool containing X number of 'workers' is defined
   • Each worker takes an ID and makes a request (with up to X concurrent workers running at a time)
   • When a worker completes, it takes the next ID from the array and processes that
   • etc. until all IDs have been processed

Any experience welcome

Answer 1

It was actually a lot simpler than I initially thought, and only requires Bluebird (I'm paraphrasing here a little bit since my final code ended up much more complex):

var Promise = require('bluebird');
...
var allResults = [];
...
Promise.map(idList, (id) => {
      // For each ID in idList, make a HTTP call
      return http.get( ... url: 'http://foo.com/bar?id=' + id ... )
              .then((httpResposne) => {
                return allResults.push(httpResposne);
              })
              .catch((err) => {
                var errMsg = 'ERROR: [' + err + ']';
                console.log(errMsg + (err.stack ? '\n' + err.stack : ''));

              });
    }, { concurrency: 10 }) // Max of 10 concurrent HTTP calls at once
    .then(() => {
      // All requests are now complete, return all results
      return res.json(allResults);
    });