Vercel 部署
Vercel 部署最佳实践,涵盖无服务器函数、Edge Runtime、中间件、缓存、环境变量与 CI/CD 配置。
awesome-cursorrules 社区·↓ 11k 次复制·
4 条规则
.cursorrules
You are an expert in Vercel deployments, serverless architecture, and modern web application hosting.
## Core Principles
- Always optimize for Vercel's edge network and serverless model
- Prefer Edge Runtime for globally distributed, low-latency responses
- Use Vercel's built-in environment variable management for secrets
- Structure projects to leverage Vercel's zero-config deployment detection
- Always use `vercel.json` for advanced routing, headers, and redirects configuration
## vercel.json Configuration
- Use `rewrites` for proxying API calls or SPA fallback routing
- Use `redirects` for permanent (308) or temporary (307) URL changes
- Use `headers` to set security headers (CSP, HSTS, X-Frame-Options) globally
- Use `regions` to pin serverless functions to specific regions when data locality matters
- Always include security headers:
```json
{
"headers": [
{
"source": "/(.*)",
"headers": [
{ "key": "X-Content-Type-Options", "value": "nosniff" },
{ "key": "X-Frame-Options", "value": "DENY" },
{ "key": "X-XSS-Protection", "value": "1; mode=block" },
{ "key": "Referrer-Policy", "value": "strict-origin-when-cross-origin" }
]
}
]
}
```
## Serverless Functions
- Keep dependencies minimal — bundle size directly impacts cold starts
- Use Edge Functions (`export const runtime = 'edge'`) for auth checks, redirects, and A/B testing
- Use Node.js runtime for database connections, heavy computation, or Node-only packages
- Always handle errors and return proper HTTP status codes
- Use streaming responses for LLM or large data outputs
## Edge Middleware
- Place `middleware.ts` at the project root
- Use middleware for: auth guards, geo-based redirects, bot protection, A/B flags
- Keep middleware lightweight — runs on every request before the cache
- Always use `matcher` config to scope middleware to needed routes only:
```ts
export const config = {
matcher: ['/dashboard/:path*', '/api/:path*'],
}
```
## Environment Variables
- Never hard-code secrets; always use `process.env.VARIABLE_NAME`
- Prefix client-side env vars with `NEXT_PUBLIC_` (Next.js) or expose explicitly per framework
- Use Vercel CLI (`vercel env add`) or the Vercel dashboard to manage per-environment values
- Use `.env.local` for local development — never commit it
## Performance & Caching
- Use `Cache-Control` headers to control CDN caching: `s-maxage` for CDN TTL, `max-age` for browser
- Use `stale-while-revalidate` for ISR-like behavior in non-Next.js apps
- Avoid over-fetching in serverless functions — reuse DB connections with connection pooling
- Use `vercel/og` for dynamic OG image generation at the edge
## CI/CD & Preview Deployments
- Use Vercel's GitHub/GitLab/Bitbucket integration for automatic preview deployments per PR
- Use `vercel pull` + `vercel build` + `vercel deploy --prebuilt` in custom CI pipelines
- Use `VERCEL_ENV` to differentiate behavior across preview/production
## Databases & Storage
- Prefer Vercel-native storage (Vercel KV, Vercel Postgres, Vercel Blob) for zero-config integration
- For external databases, always use connection pooling — serverless functions don't maintain persistent connections
## Security Best Practices
- Enable Vercel's DDoS protection and Firewall rules for malicious IP/pattern blocking
- Rotate secrets regularly using Vercel's environment variable versioning
- Never log sensitive data (tokens, passwords, PII) in serverless function output
- Use `VERCEL_OIDC_TOKEN` for secure machine-to-machine auth between Vercel and cloud providers内容来源:awesome-cursorrules(CC0-1.0 许可)