如何解决Puppeteer中net::ERR_CONNECTION_RESET等系列连接错误

First, let's break down your problem: those ERR_CONNECTION_RESET/CLOSED/FAILED errors are likely a mix of unreliable free proxies and Google flagging your Puppeteer instance as a bot. Let's start with beefing up your bot detection evasion, then cover proxy best practices.

Step 1: Enhance Your preparePageForTests Function #

Add these additional mocks and tweaks to make your browser instance look far more like a real human user:

const preparePageForTests = async (page) => { 
  const userAgent = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/78.0.3904.108 Safari/537.36'; 
  await page.setUserAgent(userAgent); 

  // Critical: Block Chrome's AutomationControlled flag (newer detection method)
  await page.evaluateOnNewDocument(() => {
    delete window.navigator.__proto__.webdriver;
    Object.defineProperty(navigator, 'webdriver', {
      get: () => undefined,
    });
  });

  // Mock real desktop hardware specs
  await page.evaluateOnNewDocument(() => {
    Object.defineProperty(navigator, 'hardwareConcurrency', {
      get: () => 8, // Typical desktop core count
    });
    Object.defineProperty(navigator, 'deviceMemory', {
      get: () => 16, // Typical desktop RAM
    });
  });

  // Set realistic viewport and screen properties
  await page.setViewport({
    width: 1920,
    height: 1080,
    deviceScaleFactor: 1,
  });
  await page.evaluateOnNewDocument(() => {
    Object.defineProperty(window.screen, 'colorDepth', {
      get: () => 24,
    });
  });

  // Mock network connection details
  await page.evaluateOnNewDocument(() => {
    Object.defineProperty(navigator, 'connection', {
      get: () => ({
        type: 'wifi',
        effectiveType: '4g',
        downlink: 10,
        rtt: 50,
      }),
    });
  });

  // Pass the Chrome Test. 
  await page.evaluateOnNewDocument(() => { 
    window.navigator.chrome = { 
      app: { isInstalled: false, }, 
      webstore: { onInstallStageChanged: {}, onDownloadProgress: {}, }, 
      runtime: { 
        PlatformOs: { MAC: 'mac', WIN: 'win', ANDROID: 'android', CROS: 'cros', LINUX: 'linux', OPENBSD: 'openbsd', }, 
        PlatformArch: { ARM: 'arm', X86_32: 'x86-32', X86_64: 'x86-64', }, 
        PlatformNaclArch: { ARM: 'arm', X86_32: 'x86-32', X86_64: 'x86-64', }, 
        RequestUpdateCheckStatus: { THROTTLED: 'throttled', NO_UPDATE: 'no_update', UPDATE_AVAILABLE: 'update_available', }, 
        OnInstalledReason: { INSTALL: 'install', UPDATE: 'update', CHROME_UPDATE: 'chrome_update', SHARED_MODULE_UPDATE: 'shared_module_update', }, 
        OnRestartRequiredReason: { APP_UPDATE: 'app_update', OS_UPDATE: 'os_update', PERIODIC: 'periodic', }, 
      } 
    }; 
  }); 

  await page.evaluateOnNewDocument(() => { 
    const originalQuery = window.navigator.permissions.query; 
    return window.navigator.permissions.query = (parameters) => ( 
      parameters.name === 'notifications' ? Promise.resolve({ state: Notification.permission }) : originalQuery(parameters) 
    ); 
  }); 

  await page.evaluateOnNewDocument(() => { 
    Object.defineProperty(navigator, 'plugins', { 
      get: () => [1, 2, 3, 4, 5], 
    }); 
  }); 

  await page.evaluateOnNewDocument(() => { 
    Object.defineProperty(navigator, 'languages', { 
      get: () => ['en-US', 'en'], 
    }); 
  }); 

  // Add random delay to mimic human hesitation before navigation
  await page.waitForTimeout(Math.floor(Math.random() * 2000) + 1000);
}

Step 2: Update Your Launch Arguments #

Add these critical flags to reduce detection and fix Heroku-specific resource issues:

const browser = await puppeteerExtra.launch({ 
  headless: 'new', // Newer headless mode is far more undetectable
  args: [ 
    '--no-sandbox', 
    '--disable-setuid-sandbox', 
    '--proxy-server=socks4://109.94.182.128:4145',
    '--disable-blink-features=AutomationControlled', // Blocks core automation detection
    '--disable-dev-shm-usage', // Fixes memory limits on Heroku
    '--disable-features=VizDisplayCompositor', // Reduces fingerprinting surface
    '--start-maximized', // Matches our realistic viewport
  ]});

Step 3: Fix Proxy Reliability #

Free SOCKS4 proxies from public lists are extremely unreliable—they get blocked by Google within minutes, are often slow, or dead on arrival. Here's how to fix this:

• Use paid rotating residential proxies: Services like BrightData, Oxylabs, or Smartproxy offer proxies that mimic real user IPs, making them far less likely to get blocked.
• Test proxies before use: Add a quick check to verify the proxy works (e.g., navigate to https://api.ipify.org and confirm the IP matches your proxy).
• Handle failures gracefully: Wrap your navigation code in try/catch blocks, and retry with a different proxy if you hit a connection error.

Step 4: Use Puppeteer Extra Stealth Plugin (Optional but Recommended) #

Since you're already using puppeteer-extra, install the stealth plugin to automate most evasion work:

npm install puppeteer-extra-plugin-stealth

Then update your code to use it:

const puppeteerExtra = require('puppeteer-extra');
const StealthPlugin = require('puppeteer-extra-plugin-stealth');

puppeteerExtra.use(StealthPlugin());

// Rest of your code can drop most manual mocks—this plugin handles them automatically

Final Tips #

• Rotate user agents: Use a list of real Chrome user agents and pick one randomly each run.
• Add human-like actions: Insert random delays between clicks, scrolls, and navigation to avoid triggering rate limits.
• Upgrade Heroku dyno: Puppeteer is resource-heavy—use at least a standard dyno to avoid memory crashes.

Content of this question originates from Stack Exchange, asked by nickcoding2