后端服务
graphql-api-caching
更新于 2026-10-28
GraphQL · 缓存规则
Caching rules for GraphQL: explicit invalidation owners, versioned keys and stampede protection.
- 技术栈
- GraphQL
- 主题
- 缓存
- 条目
- 9 条
- 版本
- v0.2.0
✓ 应当这么做
- ✓ Name the invalidation owner for every cache entry
- ✓ Version cache keys so a deploy can invalidate without a purge
- ✓ Protect against stampedes with single-flight or jittered expiry
✕ 常见反模式
- ✕ Caching data whose correctness depends on user-specific authorisation without scoping the key
- ✕ Infinite TTLs with no owner
- ✕ Treating the cache as durable storage
◎ 交付前验证
- ◎ Integration tests cover the success and the failure branch of each new route
- ◎ No unhandled promise rejection warnings in the test output
- ◎ Graceful shutdown drains in-flight requests
同分类条目
36 篇后端服务
Node.js 工程规则
Node.js async rules: validated boundaries, non-blocking I/O and graceful shutdown
后端服务
NestJS 工程规则
NestJS rules: thin controllers, validated DTOs and evented module boundaries
后端服务
Spring Boot 工程规则
Spring Boot rules: constructor injection, transaction boundaries and validated input
后端服务
FastAPI 工程规则
FastAPI rules: Pydantic boundary models and dependency-injected services
后端服务
Django 工程规则
Django rules: queryset-level filtering and explicit select_related usage
后端服务
Flask 工程规则
Flask rules: blueprint-scoped config and application-factory wiring