@foo-software/lighthouse-check

An NPM module and CLI for automated Lighthouse audits.

Usage no npm install needed!

<script type="module">
  import fooSoftwareLighthouseCheck from 'https://cdn.skypack.dev/@foo-software/lighthouse-check';
</script>

README

CircleCI

@foo-software/lighthouse-check

An NPM module and CLI to run Lighthouse audits programmatically. This project aims to add bells and whistles to automated Lighthouse testing for DevOps workflows. Easily implement in your Continuous Integration or Continuous Delivery pipeline.

This project provides two ways of running audits - locally in your own environment or remotely via Foo's Automated Lighthouse Check API. For basic usage, running locally will suffice, but if you'd like to maintain a historical record of Lighthouse audits and utilize other features, you can run audits remotely by following the steps and examples.

Features

Table of Contents

Install

npm install @foo-software/lighthouse-check

Usage

@foo-software/lighthouse-check provides several functionalities beyond standard Lighthouse audits. It's recommended to start with a basic implementation and expand on it as needed.

Basic Usage

Calling lighthouseCheck will run Lighthouse audits against https://www.foo.software/lighthouse and https://www.foo.software/contact.

import { lighthouseCheck } from '@foo-software/lighthouse-check';

(async () => {
  const response = await lighthouseCheck({
    urls: [
      'https://www.foo.software/lighthouse',
      'https://www.foo.software/contact'
    ]
  });

  console.log('response', response);
})();

Or via CLI.

$ lighthouse-check --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact"

The CLI will log the results.

lighthouse-check CLI output

Foo's Automated Lighthouse Check API Usage

Foo's Automated Lighthouse Check can monitor your website's quality by running audits automatically! It can provide a historical record of audits over time to track progression and degradation of website quality. Create a free account to get started. With this, not only will you have automatic audits, but also any that you trigger additionally. Below are steps to trigger audits on URLs that you've created in your account.

Trigger Audits on All Pages in an Account

Basic example with the CLI

$ lighthouse-check --apiToken "abcdefg"

Trigger Audits on Only Certain Pages in an Account

  • Navigate to your account details, click into "Account Management" and make note of the "API Token".
  • Navigate to your dashboard and once you've created URLs to monitor, click on the "More" link of the URL you'd like to use. From the URL details screen, click the "Edit" link at the top of the page. You should see an "API Token" on this page. It represents the token for this specific page (not to be confused with an account API token).
  • Use the account token as the apiToken option and page token (or group of page tokens) as urls option.

Basic example with the CLI

$ lighthouse-check --apiToken "abcdefg" \
  --urls "hijklmnop,qrstuv"

You can combine usage with other options for a more advanced setup. Example below.

Runs audits remotely and posts results as comments in a PR

$ lighthouse-check --apiToken "abcdefg" \
  --urls "hijklmnop,qrstuv" \
  --prCommentAccessToken "abcpersonaltoken" \
  --prCommentUrl "https://api.github.com/repos/foo-software/lighthouse-check/pulls/3/reviews"

Saving Reports Locally

You may notice above we had two lines of output; Report and Local Report. These values are populated when options are provided to save the report locally and to S3. These options are not required and can be used together or alone.

Saving a report locally example below.

import { lighthouseCheck } from '@foo-software/lighthouse-check';

(async () => {
  const response = await lighthouseCheck({
    // relative to the file. NOTE: when using the CLI `--outputDirectory` is relative
    // to where the command is being run from.
    outputDirectory: '../artifacts',
    urls: [
      'https://www.foo.software/lighthouse',
      'https://www.foo.software/contact'
    ]
  });

  console.log('response', response);
})();

Or via CLI.

$ lighthouse-check --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact" \
  --ouputDirectory "./artifacts"

Saving Reports to S3

import { lighthouseCheck } from '@foo-software/lighthouse-check';

(async () => {
  const response = await lighthouseCheck({
    awsAccessKeyId: 'abc123',
    awsBucket: 'my-bucket',
    awsRegion: 'us-east-1',
    awsSecretAccessKey: 'def456',
    urls: [
      'https://www.foo.software/lighthouse',
      'https://www.foo.software/contact'
    ]
  });

  console.log('response', response);
})();

Or via CLI.

$ lighthouse-check --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact" \
  --awsAccessKeyId abc123 \
  --awsBucket my-bucket \
  --awsRegion us-east-1 \
  --awsSecretAccessKey def456 \

Implementing with Slack

Below is a basic Slack implementation. To see how you can accomplish notifications with code versioning data - see the CircleCI example (ie GitHub authors, PRs, branches, etc).

import { lighthouseCheck } from '@foo-software/lighthouse-check';

(async () => {
  const response = await lighthouseCheck({
    slackWebhookUrl: 'https://www.my-slack-webhook-url.com',
    urls: [
      'https://www.foo.software/lighthouse',
      'https://www.foo.software/contact'
    ]
  });

  console.log('response', response);
})();

Or via CLI.

$ lighthouse-check --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact" \
  --slackWebhookUrl "https://www.my-slack-webhook-url.com"

The below screenshot shows an advanced implementation as detailed in the CircleCI example.

Lighthouse Check Slack notification

Enabling PR Comments

Populate prCommentAccessToken and prCommentUrl options to enable comments on pull requests.

Lighthouse Check PR comments

Enforcing Minimum Scores

You can use validateStatus to enforce minimum scores. This could be handy in a DevOps workflow for example.

import { lighthouseCheck, validateStatus } from '@foo-software/lighthouse-check';

(async () => {
  try {
    const response = await lighthouseCheck({
      awsAccessKeyId: 'abc123',
      awsBucket: 'my-bucket',
      awsRegion: 'us-east-1',
      awsSecretAccessKey: 'def456',
      urls: [
        'https://www.foo.software/lighthouse',
        'https://www.foo.software/contact'
      ]
    });

    const status = await validateStatus({
      minAccessibilityScore: 90,
      minBestPracticesScore: 90,
      minPerformanceScore: 70,
      minProgressiveWebAppScore: 70,
      minSeoScore: 80,
      results: response
    });

    console.log('all good?', status); // 'all good? true'
  } catch (error) {
    console.log('error', error.message);

    // log would look like:
    // Minimum score requirements failed:
    // https://www.foo.software/lighthouse: Performance: minimum score: 70, actual score: 64
    // https://www.foo.software/contact: Performance: minimum score: 70, actual score: 44
  }
})();

Or via CLI. Important: outputDirectory value must be defined and the same in both commands.

$ lighthouse-check --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact" \
  --outputDirectory /tmp/artifacts \
$ lighthouse-check-status --outputDirectory /tmp/artifacts \
  --minAccessibilityScore 90 \
  --minBestPracticesScore 90 \
  --minPerformanceScore 70 \
  --minProgressiveWebAppScore 70 \
  --minSeoScore 80

Implementing with CircleCI

In the below example we run Lighthouse audits on two URLs, save reports as artifacts, deploy reports to S3 and send a Slack notification with GitHub info. We defined environment variables like LIGHTHOUSE_CHECK_AWS_BUCKET in the CircleCI project settings.

This implementation utilizes a CircleCI Orb - lighthouse-check-orb.

version: 2.1

orbs:
  lighthouse-check: foo-software/lighthouse-check@0.0.6 # ideally later :)

jobs:
  test: 
    executor: lighthouse-check/default
    steps:
      - lighthouse-check/audit:
          urls: https://www.foo.software/lighthouse,https://www.foo.software/contact
          # this serves as an example, however if the below environment variables
          # are set - the below params aren't even necessary. for example - if
          # LIGHTHOUSE_CHECK_AWS_ACCESS_KEY_ID is already set - you don't need
          # the line below.
          awsAccessKeyId: $LIGHTHOUSE_CHECK_AWS_ACCESS_KEY_ID
          awsBucket: $LIGHTHOUSE_CHECK_AWS_BUCKET
          awsRegion: $LIGHTHOUSE_CHECK_AWS_REGION
          awsSecretAccessKey: $LIGHTHOUSE_CHECK_AWS_SECRET_ACCESS_KEY
          slackWebhookUrl: $LIGHTHOUSE_CHECK_SLACK_WEBHOOK_URL

workflows:
  test:
    jobs:
      - test
lighthouse-check CircleCI post-deploy

Reports are saved as "artifacts".

lighthouse-check CircleCI post-deploy artifacts

Upon clicking the HTML file artifacts, we can see the full report!

lighthouse-check CircleCI post-deploy artifact Lighthouse report

In the example above we also uploaded reports to S3. Why would we do this? If we want to persist historical data - we don't want to rely on temporary cloud storage.

Implementing with GitHub Actions

Similar to the CircleCI implementation, we can also create a workflow implementation with GitHub Actions using lighthouse-check-action. Example below.

.github/workflows/test.yml

name: Test Lighthouse Check
on: [push]

jobs:
  lighthouse-check:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@master
      - run: mkdir /tmp/artifacts
      - name: Run Lighthouse
        uses: foo-software/lighthouse-check-action@master
        with:
          accessToken: ${{ secrets.LIGHTHOUSE_CHECK_GITHUB_ACCESS_TOKEN }}
          author: ${{ github.actor }}
          awsAccessKeyId: ${{ secrets.LIGHTHOUSE_CHECK_AWS_ACCESS_KEY_ID }}
          awsBucket: ${{ secrets.LIGHTHOUSE_CHECK_AWS_BUCKET }}
          awsRegion: ${{ secrets.LIGHTHOUSE_CHECK_AWS_REGION }}
          awsSecretAccessKey: ${{ secrets.LIGHTHOUSE_CHECK_AWS_SECRET_ACCESS_KEY }}
          branch: ${{ github.ref }}
          outputDirectory: /tmp/artifacts
          urls: 'https://www.foo.software/lighthouse,https://www.foo.software/contact'
          sha: ${{ github.sha }}
          slackWebhookUrl: ${{ secrets.LIGHTHOUSE_CHECK_WEBHOOK_URL }}
      - name: Upload artifacts
        uses: actions/upload-artifact@master
        with:
          name: Lighthouse reports
          path: /tmp/artifacts

Overriding Config and Option Defaults

You can override default config and options by specifying overridesJsonFile option. Contents of this overrides JSON file can have two possible fields; options and config. These two fields are eventually used by Lighthouse to populate opts and config arguments respectively as illustrated in Using programmatically. The two objects populating this JSON file are merged shallowly with the default config and options.

Example content of overridesJsonFile

{
  "config": {
    "settings": {
      "onlyCategories": ["performance"]
    }
  },
  "options": {
    "chromeFlags": [
      "--disable-dev-shm-usage"
    ]
  }
}

CLI

Running lighthouse-check in the example below will run Lighthouse audits against https://www.foo.software/lighthouse and https://www.foo.software/contact and output a report in the '/tmp/artifacts' directory.

Format is --option <argument>. Example below.

$ lighthouse-check --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact" \
  --outputDirectory /tmp/artifacts

lighthouse-check-status example

$ lighthouse-check-status --outputDirectory /tmp/artifacts \
  --minAccessibilityScore 90 \
  --minBestPracticesScore 90 \
  --minPerformanceScore 70 \
  --minProgressiveWebAppScore 70 \
  --minSeoScore 80

CLI Options

All options mirror the NPM module. The only difference is that array options like urls are passed in as a comma-separated string as an argument using the CLI.

Docker

$ docker pull foosoftware/lighthouse-check:latest
$ docker run foosoftware/lighthouse-check:latest \
  lighthouse-check --verbose \
  --urls "https://www.foo.software/lighthouse,https://www.foo.software/contact"

Options

lighthouse-check functions accept a single configuration object.

lighthouseCheck

You can choose from two ways of running audits - locally in your own environment or remotely via Foo's Automated Lighthouse Check API. You can think of local runs as the default implementation. For directions about how to run remotely see the Foo's Automated Lighthouse Check API Usage section. We denote which options are available to a run type with the Run Type values of either local, remote, or both.

Below are options for the exported lighthouseCheck function or lighthouse-check command with CLI.

Name Description Type Run Type Default Required
apiToken The foo.software account API token found in the dashboard. string remote undefined no
author For Slack notifications: A user handle, typically from GitHub. string both undefined no
awsAccessKeyId The AWS accessKeyId for an S3 bucket. string local undefined no
awsBucket The AWS Bucket for an S3 bucket. string local undefined no
awsRegion The AWS region for an S3 bucket. string local undefined no
awsSecretAccessKey The AWS secretAccessKey for an S3 bucket. string local undefined no
branch For Slack notifications: A version control branch, typically from GitHub. string both undefined no
configFile A configuration file path in JSON format which holds all options defined here. This file should be relative to the file being interpretted. string both undefined no
extraHeaders HTTP Header key/value pairs to send in requests. If using the CLI this will need to be stringified, for example: "{\"Cookie\":\"monster=blue\", \"x-men\":\"wolverine\"}" object local undefined no
emulatedFormFactor Lighthouse setting only used for local audits. See src/lighthouseConfig.js comments for details. oneOf(['mobile', 'desktop', 'all']) local undefined no
locale A locale for Lighthouse reports. Example: ja string local undefined no
maxRetries The maximum number of times to retry.Note: This is not supported when running against Foo's API as retry logic is already in place. number local 0 no
maxWaitForLoad The maximum amount of time to wait for a page to load in ms. number local undefined no
overridesJsonFile A JSON file with config and option fields to overrides defaults. Read more here. string local undefined no
outputDirectory An absolute directory path to output report. You can do this an an alternative or combined with an S3 upload. string local undefined no
pr For Slack notifications: A version control pull request URL, typically from GitHub. string local undefined no
prCommentAccessToken Access token of a user to post PR comments. string both undefined no
prCommentEnabled If true and prCommentAccessToken is set along with prCommentUrl, scores will be posted as comments. boolean both true no
prCommentSaveOld If true and PR comment options are set, new comments will be posted on every change vs only updating once comment with most recent scores. boolean both false no
prCommentUrl An endpoint to post comments to. Typically this will be from GitHub's API. Example: https://api.github.com/repos/:owner/:repo/pulls/:pull_number/reviews string both undefined no
slackWebhookUrl A Slack Incoming Webhook URL to send notifications to. string both undefined no
sha For Slack notifications: A version control sha, typically from GitHub. string both undefined no
tag An optional tag or name (example: build #2 or v0.0.2). string remote undefined no
throttlingMethod Lighthouse setting only used for local audits. See src/lighthouseConfig.js comments for details. oneOf(['simulate', 'devtools', 'provided']) local undefined no
throttling Lighthouse setting only used for local audits. See src/lighthouseConfig.js comments for details. oneOf(['mobileSlow4G', 'mobileRegluar3G', 'desktopDense4G']) local undefined no
timeout Minutes to timeout. If wait is true (it is by default), we wait for results. If this timeout is reached before results are received an error is thrown. number local 10 no
urls An array of URLs (or page API tokens if running remotely). In the CLI this value should be a comma-separated list. array both undefined yes
verbose If true, print out steps and results to the console. boolean both true no
wait If true, waits for all audit results to be returned, otherwise URLs are only enqueued. boolean remote true no

validateStatus

results parameter is required or alternatively outputDirectory. To utilize outputDirectory - the same value would also need to be specified when calling lighthouseCheck.

Name Description Type Default Required
minAccessibilityScore The minimum accessibility Lighthouse score required. number undefined no
minBestPracticesScore The minimum best practices Lighthouse score required. number undefined no
minPerformanceScore The minimum performance Lighthouse score required. number undefined no
minProgressiveWebAppScore The minimum progressive web app Lighthouse score required. number undefined no
minSeoScore The minimum SEO Lighthouse score required. number undefined no
outputDirectory An absolute directory path to output report. When the results object isn't specified, this value will need to be. string undefined no
results A results object representing results of Lighthouse audits. number undefined no

Return Payload

lighthouseCheck function returns a promise which either resolves as an object or rejects as an error object. In both cases the payload will be of the same shape documented below.

Name Description Type
code A code to signify failure or succes. oneOf(["SUCCESS", "ERROR_GENERIC", ...]) see errorCodes.js for all error codes.
data An array of results returned by the API. array
message A message to elaborate on the code. This field isn't always populated. string

Credits

This package was brought to you by Foo - a website performance monitoring tool. Create a free account with standard performance testing. Automatic website performance testing, uptime checks, charts showing performance metrics by day, month, and year. Foo also provides real time notifications when performance and uptime notifications when changes are detected. Users can integrate email, Slack and PagerDuty notifications.