ขนาดชุดรวมและการเพิ่มประสิทธิภาพโค้ด

Workers have a script size limit (1 MB compressed on the free plan, more on paid).

Smaller bundles mean:

• Faster parse and startup
• Lower memory footprint
• Reduced cold-start impact

Optimizing your code is a direct performance win at the edge.

Wrangler reports your bundle size on every build. Always measure first.

Use ES module import/export so the bundler can tree-shake unused code.

Import only what you need, never the whole library when a single function will do.

Heavy npm packages bloat bundles fast. Prefer small, edge-friendly libraries.

• Use the platform URL, crypto.subtle, and fetch instead of polyfills
• Swap moment.js for a tiny date helper
• Audit with a bundle analyzer

Workers and Deno expose many Web Platform APIs natively, so you avoid bundling polyfills.

Wrangler minifies production builds by default, but confirm it is enabled.

Workers bill and limit by CPU time, not wall-clock. Optimize hot paths:

• Avoid synchronous loops over huge arrays
• Cache computed results in KV or memory
• Stream large responses instead of buffering

Streaming sends data as it is produced, keeping memory low and CPU steady.

Dynamic import() can defer loading code paths that are not always needed, keeping the initial module lean.

Use an analyzer to see which dependencies dominate. esbuild (which Wrangler uses) can emit a metafile.

To keep edge code fast and small:

• Measure bundle size on every deploy
• Tree-shake and import narrowly
• Prefer native Web APIs over polyfills
• Minify, stream, and lazy-load
• Cut CPU time in hot paths

Which technique most directly lets the bundler remove unused library code?

You optimized both size and speed:

• Measure first, then tree-shake and minify
• Use native Web APIs and lightweight deps
• Stream and lazy-load to cut memory and CPU
• Analyze the bundle to find offenders

Lean Workers start faster and cost less, every kilobyte counts at the edge.