caching-strategy

Files (2)

• README.md
• SKILL.md

SKILL.md (first lines)

# Caching Strategy

## Rules

- Cache at every layer: browser (Cache-Control), CDN (edge), application (Redis), database (query cache)
- Use `stale-while-revalidate` for content that can be slightly stale — serve fast, refresh in background
- Cache-Control headers: `public, max-age=31536000, immutable` for hashed static assets; `public, s-maxage=60, stale-while-revalidate=300` for dynamic pages
- Redis caching: serialize to JSON, set TTL on every key, use namespaced keys (`cache:users:123`)
- React Query / TanStack Query: set `staleTime` (how long data is fresh) and `gcTime` (how long to keep in memory)
- Invalidate on write — when data changes, delete or update the cache entry immediately
- Use cache-aside pattern: check cache first, fetch from source on miss, write to cache
- For Next.js: leverage ISR (`revalidate`), `unstable_cache`, and fetch cache for server components
- Tag-based invalidation: tag cached entries by entity type, invalidate all entries for a tag on mutation

## Patterns

```ts
// Cache-aside with Redis
async function getCachedUser(id: string) {
  const cached = await redis.get(`user:${id}`);
  if (cached) return JSON.parse(cached);
  const user = await db.user.findUnique({ where: { id } });
  await redis.set(`user:${id}`, JSON.stringify(user), "EX", 3600);
  return user;
}
```

## Avoid

- Caching without TTL — stale data will persist indefinitely
- Caching user-specific data in shared CDN cache — use `private` or `Vary` headers
- Cache stampede: use locking or `stale-while-revalidate` to prevent thundering herd
- Caching error responses — only cache successful results

From YepAPI/skills, MIT.

More from YepAPI/skills