Performance & Optimization
Core Topics & Concepts
1. Large List Optimization (100k+ Items)
- Virtualization: Use
react-windoworreact-virtualizedto render only items currently visible in the viewport plus a small overscan buffer. - Pagination & Infinite Scroll: Load data lazily in chunks via cursor-based API requests.
- Row Stability: Use stable keys and fixed/predictable row heights to prevent layout recalculation during scrolling.
2. Global Load Time Optimization
- CDN & Edge Caching: Distribute static assets close to users worldwide.
- Compression: Use Brotli / Gzip compression for text-based assets.
- Code Splitting: Dynamic imports (
import()) and route-based code splitting. - Image Optimization: WebP/AVIF formats, responsive
srcset, lazy loading, and reserving width/height to avoid layout shifts. - Preconnect & Preload: Preconnect to critical domains and preload critical LCP assets.
- SSR / SSG: Pre-render pages for rapid first contentful paint.
3. Preventing Main Thread Blocking
- Web Workers: Move CPU-heavy data parsing, filtering, and sorting off the main UI thread.
- Time Slicing: Chunk heavy operations using
requestIdleCallbackor scheduler timeouts. - Concurrent React: Use
useTransitionto keep user inputs responsive during heavy renders.
4. DevTools Debugging Workflow
- React Profiler: Capture render durations, flame charts, and turn on “Highlight updates when components render” to spot unnecessary renders.
- Chrome Performance Tab: Identify long tasks (>50ms), main-thread blocking JavaScript, and layout thrashing.
5. Smooth & Unjanky Animations
- Compositor Properties: Animate only GPU-accelerated properties (
transformandopacity). - Avoid Layout / Reflow: Do not animate properties that trigger layout (
width,height,top,left,margin). - DevTools: Inspect with the Rendering tab to check Paint Flashing and FPS Meter.
6. Core Web Vitals (CWV)
- LCP (Largest Contentful Paint): Loading performance. Optimize LCP by preloading hero images and serving critical CSS inline.
- INP (Interaction to Next Paint): Interactivity & responsiveness. Reduce main thread blocking and JavaScript execution time.
- CLS (Cumulative Layout Shift): Visual stability. Always set explicit
widthandheightdimensions on images/iframes and reserve space for dynamic ads.
7. Bundle Optimization
- Tree Shaking: Ensure ES modules format (
import/export) for effective dead-code elimination. - Bundle Analyzer: Inspect bundle output using
webpack-bundle-analyzerorrollup-plugin-visualizer. - Library Replacement: Replace bloated dependencies (e.g., Moment.js -> date-fns / Day.js).
Top Interview Questions
Q1. Optimize a React app rendering 100k+ list items
Explanation
Rendering 100k DOM nodes freezes the browser. The interviewer wants to see that you know the DOM, not React, is the bottleneck.
Answer
- Virtualize the list (
react-window/react-virtualized) so only visible rows plus a small buffer are rendered. - Paginate or infinite-scroll if the data comes from an API.
- Use stable keys and
React.memoon row components. - Keep row height fixed or measured so scrolling stays smooth.
- Filter and sort on the server or in a Web Worker, not during render.
Q2. Improve page load time for a global audience
Explanation
Tests whether you think beyond code: network, geography, and caching.
Answer
- Serve static assets from a CDN with edge caching.
- Compress with Brotli, and use HTTP/2 or 3.
- Code split and lazy-load routes.
- Optimize images (WebP/AVIF, srcset, lazy loading).
- Use SSR/SSG for a fast first paint.
- Preconnect to critical origins and preload the LCP resource.
- Measure with real-user monitoring per region.
Q5. Debug a React performance bottleneck using DevTools
Explanation
They want a concrete workflow, not “I’d optimize it.”
Answer
- Record in the React Profiler and find components that re-render often or slowly. Turn on “why did this render.”
- Record in the Chrome Performance tab to find long tasks, layout thrashing, and expensive scripting.
- Fix the cause: memoize, colocate state, split contexts, virtualize, or defer work.
- Re-profile to confirm the improvement.
Q12. Render a large dataset without blocking the main thread
Explanation
Tests your understanding of the single-threaded browser.
Answer
- Move heavy computation to a Web Worker.
- Break work into chunks using
requestIdleCallbackor scheduling. - Virtualize rendering.
- Use
useTransition/useDeferredValueto keep input responsive. - Paginate or stream data from the server.
Q14. Janky CSS animation on mobile
Explanation
Tests your knowledge of the browser rendering pipeline.
Answer
- Identify: Use the DevTools Performance and Rendering tabs (paint flashing, FPS meter) and test on a real low-end device.
- Fix: Animate only
transformandopacity, which the compositor handles on the GPU. Avoid animatingwidth,height,top, orleft, which trigger layout. Usewill-changesparingly. Reduce paint area and layer count. Respectprefers-reduced-motion.