Imported from jeremylongshore/tons-of-skills-marketplace (skills/.curated/vercel-performance-tuning/SKILL.md). Install upstream with npx skills add jeremylongshore/tons-of-skills-marketplace --skill vercel-performance-tuning. Copyright stays with the author (MIT).

Vercel Performance Tuning

Overview

Optimize Vercel deployment performance across four levers: edge caching, bundle size reduction, serverless function cold start elimination, and Core Web Vitals improvement. Uses real Vercel cache headers, ISR, and Edge Functions for maximum performance.

Prerequisites

• Vercel project deployed with accessible URL
• Access to Vercel Analytics (dashboard)
• Bundle analyzer available (@next/bundle-analyzer or similar)

Instructions

Step 1: Establish Performance Baseline

# Check deployment size and function count
vercel inspect https://my-app.vercel.app

# Run Lighthouse via CLI
npx lighthouse https://my-app.vercel.app --output=json --quiet \
  | jq '{performance: .categories.performance.score, lcp: .audits["largest-contentful-paint"].numericValue, cls: .audits["cumulative-layout-shift"].numericValue}'

# Check bundle size (Next.js)
ANALYZE=true npx next build
# Opens bundle analyzer report in browser

Enable Vercel Analytics in the dashboard under Analytics tab for ongoing monitoring.

Step 2: Configure Edge Caching

// api/cached-data.ts — cache API responses at the edge
import type { VercelRequest, VercelResponse } from '@vercel/node';

export default function handler(req: VercelRequest, res: VercelResponse) {
  // Cache at Vercel edge for 60s, serve stale for 300s while revalidating
  res.setHeader('Cache-Control', 's-maxage=60, stale-while-revalidate=300');
  res.json({ data: fetchData(), cachedAt: new Date().toISOString() });
}

// vercel.json — cache static assets aggressively
{
  "headers": [
    {
      "source": "/static/(.*)",
      "headers": [
        { "key": "Cache-Control", "value": "public, max-age=31536000, immutable" }
      ]
    },
    {
      "source": "/api/public-data",
      "headers": [
        { "key": "Cache-Control", "value": "s-maxage=3600, stale-while-revalidate=86400" }
      ]
    }
  ]
}

Cache header reference:

Step 3: Incremental Static Regeneration (ISR)

// app/products/[id]/page.tsx (Next.js App Router)
export const revalidate = 60; // Revalidate every 60 seconds

export default async function ProductPage({ params }) {
  const product = await fetchProduct(params.id);
  return <ProductView product={product} />;
}

// Generate static pages at build time, regenerate on-demand
export async function generateStaticParams() {
  const products = await fetchTopProducts(100);
  return products.map(p => ({ id: p.id }));
}

On-demand revalidation via API route:

// api/revalidate.ts
import type { VercelRequest, VercelResponse } from '@vercel/node';

export default async function handler(req: VercelRequest, res: VercelResponse) {
  const secret = req.query.secret;
  if (secret !== process.env.REVALIDATION_SECRET) {
    return res.status(401).json({ error: 'Invalid secret' });
  }

  const path = req.query.path as string;
  await res.revalidate(path);
  res.json({ revalidated: true, path });
}
// Trigger: POST /api/revalidate?secret=xxx&path=/products/123

Step 4: Reduce Cold Starts

// Lazy initialization — don't import heavy modules at top level
// BAD: Cold start loads everything
import { PrismaClient } from '@prisma/client';
const prisma = new PrismaClient(); // Runs on every cold start

// GOOD: Lazy singleton — only connects when first used
let prisma: PrismaClient | null = null;
function getDb(): PrismaClient {
  if (!prisma) {
    prisma = new PrismaClient();
  }
  return prisma;
}

export default async function handler(req, res) {
  const users = await getDb().user.findMany();
  res.json(users);
}

Move latency-critical paths to Edge Functions (zero cold starts):

// api/fast.ts
export const config = { runtime: 'edge' };

export default function handler(request: Request) {
  return Response.json({ fast: true }); // No cold start, runs globally
}

Step 5: Bundle Size Optimization

// next.config.js — tree-shaking and optimization
module.exports = {
  experimental: {
    optimizePackageImports: ['lodash', '@mui/material', '@mui/icons-material'],
  },
  // Exclude server-only deps from client bundle
  webpack: (config, { isServer }) => {
    if (!isServer) {
      config.resolve.fallback = { fs: false, net: false, tls: false };
    }
    return config;
  },
};

# Find large dependencies
npx depcheck
npx cost-of-modules

# Replace heavy libraries with lighter alternatives
# moment.js (300KB) → dayjs (2KB)
# lodash (72KB) → lodash-es with tree-shaking
# axios (29KB) → native fetch

Step 6: Image Optimization

// Use Vercel's built-in image optimization
import Image from 'next/image';

// Automatic: resizes, converts to WebP/AVIF, caches at edge
<Image
  src="/hero.jpg"
  width={1200}
  height={600}
  alt="Hero"
  priority  // Preload for LCP
  sizes="(max-width: 768px) 100vw, 1200px"
/>

// vercel.json — configure image optimization
{
  "images": {
    "sizes": [640, 750, 828, 1080, 1200],
    "domains": ["images.example.com"],
    "formats": ["image/avif", "image/webp"],
    "minimumCacheTTL": 86400
  }
}

Performance Budget Reference

Output

• Edge caching configured with optimal cache-control headers
• ISR or on-demand revalidation for dynamic pages
• Cold starts eliminated via lazy initialization and Edge Functions
• Bundle size reduced through tree-shaking and import optimization
• Image optimization configured

Error Handling

Examples

Tune caching with a correctness guardrail

Choose a read-only route with no per-user content, deploy a preview change that adds a conservative s-maxage and stale-while-revalidate policy, and compare cache headers, response content, and P95 latency with synthetic requests. Keep an automated assertion that authenticated or cookie-dependent responses are never cached publicly. If the content is stale or personalized, revert the header change immediately and preserve the preview metrics for the next design review.

Resources

• Vercel Caching
• ISR Documentation
• Vercel Analytics
• Image Optimization
• Function Configuration

Next Steps

For cost optimization, see vercel-cost-tuning.