I recently rolled out a Lighthouse performance optimization checklist across the team and picked up a lot of experience along the way. I'm sharing it here in case it helps others doing similar work.

Core Concepts

Let's look at the basic implementation first:

const observer = new PerformanceObserver((list) => {
  for (const entry of list.getEntries()) {
    if (entry.entryType === 'largest-contentful-paint') {
      reportMetric('LCP', entry.startTime)
    }
    if (entry.entryType === 'first-input') {
      reportMetric('FID', entry.processingStart - entry.startTime)
    }
  }
})
observer.observe({ entryTypes: ['largest-contentful-paint', 'first-input'] })

This snippet shows the basic usage. In real projects you'll also need to consider error handling and edge cases.

In-Depth Analysis

Building on this, we can optimize further:

const observer = new PerformanceObserver((list) => {
  for (const entry of list.getEntries()) {
    if (entry.entryType === 'largest-contentful-paint') {
      reportMetric('LCP', entry.startTime)
    }
    if (entry.entryType === 'first-input') {
      reportMetric('FID', entry.processingStart - entry.startTime)
    }
  }
})
observer.observe({ entryTypes: ['largest-contentful-paint', 'first-input'] })

This pattern is very useful in large projects and can significantly cut maintenance cost.

Implementation Experience

Usage in real projects is a bit more complex:

const observer = new PerformanceObserver((list) => {
  for (const entry of list.getEntries()) {
    if (entry.entryType === 'largest-contentful-paint') {
      reportMetric('LCP', entry.startTime)
    }
    if (entry.entryType === 'first-input') {
      reportMetric('FID', entry.processingStart - entry.startTime)
    }
  }
})
observer.observe({ entryTypes: ['largest-contentful-paint', 'first-input'] })

This approach improves both the testability and extensibility of the code.

Tuning Strategies

Here is a complete example:

const observer = new PerformanceObserver((list) => {
  for (const entry of list.getEntries()) {
    if (entry.entryType === 'largest-contentful-paint') {
      reportMetric('LCP', entry.startTime)
    }
    if (entry.entryType === 'first-input') {
      reportMetric('FID', entry.processingStart - entry.startTime)
    }
  }
})
observer.observe({ entryTypes: ['largest-contentful-paint', 'first-input'] })

Pay close attention to edge cases — this is critical in production.

Summary

• A Lighthouse performance checklist is no silver bullet; choose it based on your project's size and tech stack.
• Understanding the underlying principles matters more than memorizing APIs.
• Be sure to verify compatibility before using it in production.