Debug Failed Puppeteer Tests

A guide to using the various debugging tools available in the BrowserStack Automate dashboard to debug your failed cross-browser Puppeteer tests

Interactive Session

The interactive session feature helps you take control of your Puppeteer test while it is executing on the remote browser.

You can start the interactive session by clicking on Start an interactive session link that appears in the video recording area as you execute your test script. Refer the following screenshot:

To take control of the session in the interactive mode at specific points you can add breakpoints in the test script as mentioned below:

1. Add breakpoints in the test script so that execution does not continue.
2. Execute the test, you will find the session running on the BrowserStack Automate dashboard.
3. Click on the Start an interactive session link that appears in the video recording area. (refer to the screenshot above).
4. Once the interactive session starts, you will be able to observe and take control of the ongoing test execution in the remote browser.
5. Note, that your test script should have a breakpoint else the execution will keep on happening as per the new Puppeteer commands that keep on reaching BrowserStack.

Video Recording

Every Puppeteer test run on BrowserStack is recorded exactly as it is executed on our remote machine. Video recordings are particularly helpful whenever a browser test fails as they help you retrace the steps which led to the failure. You can access the video recordings from Automate Dashboard for each session. You can also download the videos from the Dashboard or retrieve a link to download the video using our getSessionDetails executor. Learn more about getSessionDetails executor

You can run the following script at any point in your Puppeteer script to retrieve the video link using getSessionDetails:

const resp = await JSON.parse(await page.evaluate(_ => {}, `browserstack_executor: ${JSON.stringify({action: 'getSessionDetails'})}`));
console.log(resp.video_url);  // This gives the video URL of the running session. You can access it once after the test session is completed

Text Logs

Text Logs are a comprehensive record of your Puppeteer test. They are used to identify all the steps executed in the test and troubleshoot errors for the failed step. You can get verbose details of every step performed within your Puppeteer test using Test Logs.

You can access Text Logs in a session on Automate Dashboard.

You can also get the URL for downloading the Text Logs using getSessionDetails from within the test script as shown below:

const resp = await JSON.parse(await page.evaluate(_ => {}, `browserstack_executor: ${JSON.stringify({action: 'getSessionDetails'})}`));
console.log(resp.logs);  // This gives the Text Logs URL of the running session. Hit the URL once after the session is complete.

Visual Logs

Visual Logs automatically capture screenshots at every Puppeteer command executed during your test. They help debug issues detected through your Puppeteer tests. Visual Logs help with debugging the exact step and the page where the failure occurred. They also help identify any layout or design related issues with your web pages on different browsers.

Visual Logs are disabled by default. In order to enable Visual Logs you will need to set browserstack.debug capability to true.

Access Visual Logs from Automate Dashboard:

Network Logs

You can get a comprehensive log of all network activity during your Puppeteer tests using the Network Logs feature. Network Logs capture the browser’s performance data such as network traffic, latency, HTTP requests and responses in the HAR(HTTP Archive) format. You can leverage Network logs to get a clear picture of all the requests your web page is making, and the time taken for each of these requests. You can download network logs using the getSessionDetails executor or via the Automate Dashboard. You can visualize HAR files using the HAR Viewer.

Network Logs are disabled by default in your Puppeteer tests. To enable Network Logs use the capability browserstack.networkLogs with the value true within the caps JSON as shown in the sample script.

You can also get the URL for downloading the Network Logs using getSessionDetails from within the test script as shown below:

const resp = await JSON.parse(await page.evaluate(_ => {}, `browserstack_executor: ${JSON.stringify({action: 'getSessionDetails'})}`));
console.log(resp.har_logs_url);  // This gives the Network Logs URL of the running session. You can get the data from the URL after the session is complete.

Sample script

This sample script shows the usage of all the capabilities that have been explained in this document:

'use strict';
const { strict } = require('once');
const puppeteer = require('puppeteer');
const expect = require('chai').expect;
(async () => {
  const caps = {
    'browser': 'chrome',  // You can choose `chrome`, `edge` or `firefox` in this capability
    'browser_version': 'latest',  // We support v83 and above. You can choose `latest`, `latest-beta`, `latest-1`, `latest-2` and so on, in this capability
    'os': 'os x',
    'os_version': 'big sur',
    'build': 'puppeteer-build-1',  
    'name': 'My first Puppeteer test',  // The name of your test and build. See browserstack.com/docs/automate/puppeteer/organize tests for more details
    'browserstack.username': process.env.BROWSERSTACK_USERNAME || 'YOUR_USERNAME',
    'browserstack.accessKey': process.env.BROWSERSTACK_ACCESS_KEY || 'YOUR_ACCESS_KEY',
    
    'browserstack.debug': 'true',  // enabling visual logs
    'browserstack.networkLogs': 'true'  // Enabling network logs for the test
  };
  const browser = await puppeteer.connect({
  browserWSEndpoint:
  `wss://cdp.browserstack.com?caps=${encodeURIComponent(JSON.stringify(caps))}`,  // The BrowserStack CDP endpoint gives you a `browser` instance based on the `caps` that you specified
  });

  /* 
  *  The BrowserStack specific code ends here. Following this line is your test script.
  *  Here, we have a simple script that opens google.com, searches for the word BrowserStack and asserts the result.
  */
  const page = await browser.newPage();
  await page.goto('https://www.google.com/ncr');
  const element = await page.$('[aria-label="Search"]');
  await element.click();
  await element.type('BrowserStack');
  await element.press('Enter');
  const title = await page.title('');
  console.log(title);
  try {
    expect(title).to.equal("BrowserStack - Google Search", 'Expected page title is incorrect!');
    // following line of code is responsible for marking the status of the test on BrowserStack as 'passed'. You can use this code in your after hook after each test
    await page.evaluate(_ => {}, `browserstack_executor: ${JSON.stringify({action: 'setSessionStatus',arguments: {status: 'passed',reason: 'Title matched'}})}`);
  } catch {
    await page.evaluate(_ => {}, `browserstack_executor: ${JSON.stringify({action: 'setSessionStatus',arguments: {status: 'failed',reason: 'Title did not match'}})}`);
  }
  await browser.close();  // At the end of each of your tests, you must close the browser so that BrowserStack knows when to end the session.
})();