AutoGPT/autogpt_platform/frontend/LIGHTHOUSE.md

Lighthouse CI Setup

This document describes the Google Lighthouse CI integration for the AutoGPT Platform frontend.

Overview

Google Lighthouse is automatically run on the CI for every push to master or dev branches and on pull requests. It audits the application's performance, accessibility, SEO, and best practices.

Configuration

Main Configuration (lighthouserc.json)

• URLs audited: /, /build, /login
• Number of runs: 3 per URL (for consistent results)
• Thresholds:
  • Performance: 70% (warn)
  • Accessibility: 90% (error)
  • Best Practices: 85% (warn)
  • SEO: 85% (warn)
  • PWA: Disabled

Local Testing (lighthouserc.local.json)

• Lower thresholds for development testing
• Single run per URL for faster feedback

Usage

CI Integration

Lighthouse runs automatically in the lighthouse job of the platform-frontend-ci.yml workflow. Reports are uploaded as CI artifacts.

Local Development

# Start the development server
pnpm dev

# In another terminal, run Lighthouse (with more lenient thresholds)
pnpm lighthouse:local

Production Testing

# Build and start the application
pnpm build && pnpm start

# In another terminal, run full Lighthouse audit
pnpm lighthouse

Reports

• CI: Reports are stored as artifacts and can be downloaded from the GitHub Actions run page
• Local: Reports are saved to lhci_reports/ directory
• Format: Both HTML and JSON reports are generated

Customization

To modify the audit configuration:

1. URLs: Edit the collect.url array in lighthouserc.json
2. Thresholds: Adjust the assert.assertions values
3. Categories: Add or remove audit categories as needed

Troubleshooting

Common Issues

1. Frontend not ready: Ensure the application is fully started before running Lighthouse
2. Memory issues: Lighthouse can be memory-intensive; ensure sufficient resources
3. Network timeouts: Check that all services are healthy before running audits

Debug Commands

# Check configuration validity
pnpm lhci healthcheck

# Run only collection (no assertions)
pnpm lhci collect

# Run only assertions on existing reports
pnpm lhci assert

Performance Guidelines

• Performance scores below 70% will generate warnings
• Accessibility scores below 90% will fail the CI
• Focus on critical user paths for consistent performance
• Consider performance budgets for key metrics

Integration with Development Workflow

1. Pre-merge: Lighthouse runs on all PRs to catch regressions
2. Post-merge: Results on master/dev establish performance baselines
3. Monitoring: Track performance trends over time via CI artifacts