负载测试

Since Socket.IO has its own protocol, including handshake, heartbeats and custom packet encoding, the easiest way to load test your Socket.IO server is to use the Socket.IO client library and create a lot of clients.

There are two classic solutions to do this:

• using Artillery
• or manually manage the clients

Artillery

Artillery is a great tool for load testing your application. It allows creating connections, sending events and checking acknowledgments.

The documentation can be found here.

Important note: the default installation comes with a v2 client, which is not compatible with a v3/v4 server. You need to install a custom engine for this: https://github.com/ptejada/artillery-engine-socketio-v3

Installation:

$ npm install artillery artillery-engine-socketio-v3

Sample scenario:

# my-scenario.yml
config:
  target: "http://localhost:3000"
  phases:
    - duration: 60
      arrivalRate: 10
  engines:
   socketio-v3: {}

scenarios:
  - name: My sample scenario
    engine: socketio-v3
    flow:
      # wait for the WebSocket upgrade (optional)
      - think: 1

      # basic emit
      - emit:
          channel: "hello"
          data: "world"

      # emit an object
      - emit:
          channel: "hello"
          data:
            id: 42
            status: "in progress"
            tags:
              - "tag1"
              - "tag2"

      # emit in a custom namespace
      - namespace: "/my-namespace"
        emit:
          channel: "hello"
          data: "world"

      # emit with acknowledgement
      - emit:
          channel: "ping"
        acknowledge:
          match:
            value: "pong"

      # do nothing for 30 seconds then disconnect
      - think: 30

To run this scenario:

$ npx artillery run my-scenario.yml

Artillery also comes with a lot of awesome features, like the ability to publish the metrics to various endpoints or run the tests from AWS.

Its only limitation is that you cannot easily test server-to-client events, as the Artillery DSL is more suited for classic client-to-server communication. Which brings us to our next section.

Manual client creation

Here's a basic script to create a thousand Socket.IO clients and monitor the number of packets received per second:

const { io } = require("socket.io-client");

const URL = process.env.URL || "http://localhost:3000";
const MAX_CLIENTS = 1000;
const POLLING_PERCENTAGE = 0.05;
const CLIENT_CREATION_INTERVAL_IN_MS = 10;
const EMIT_INTERVAL_IN_MS = 1000;

let clientCount = 0;
let lastReport = new Date().getTime();
let packetsSinceLastReport = 0;

const createClient = () => {
  // for demonstration purposes, some clients stay stuck in HTTP long-polling
  const transports =
    Math.random() < POLLING_PERCENTAGE ? ["polling"] : ["polling", "websocket"];

  const socket = io(URL, {
    transports,
  });

  setInterval(() => {
    socket.emit("client to server event");
  }, EMIT_INTERVAL_IN_MS);

  socket.on("server to client event", () => {
    packetsSinceLastReport++;
  });

  socket.on("disconnect", (reason) => {
    console.log(`disconnect due to ${reason}`);
  });

  if (++clientCount < MAX_CLIENTS) {
    setTimeout(createClient, CLIENT_CREATION_INTERVAL_IN_MS);
  }
};

createClient();

const printReport = () => {
  const now = new Date().getTime();
  const durationSinceLastReport = (now - lastReport) / 1000;
  const packetsPerSeconds = (
    packetsSinceLastReport / durationSinceLastReport
  ).toFixed(2);

  console.log(
    `client count: ${clientCount} ; average packets received per second: ${packetsPerSeconds}`
  );

  packetsSinceLastReport = 0;
  lastReport = now;
};

setInterval(printReport, 5000);

You can use it as a starting point for load testing your own application.