Redis热键处理方案

1、分析业务场景中，QPS较大的点，并使用到的具体数据库表清单及数据特征

2、在部署Redis服务器时，建议采用一主多从或甚至多主多从架构。这种架构有助于在流量激增时，将查询请求分散到多个从库，从而有效分担负载。

3、定期监控 Redis 热键情况，发现后及时预警。

4、在应用架构层面设计多级缓存（一级会话缓存Spring Session，二级本地缓存，三级Redis缓存），以便减少Redis热键形成的概率

5、在应用中不要使用相似键，当存在大量相似键会增加哈希冲突概率，批量操作效率也会降低，因为需要单独解析执行，尤其是键大小不一时，会造成内存分配不连续，产生内存碎片

6、在应用启动时则进行热点数据预先加载，防止大并发，穿透缓存，形成雪崩

========================================

Business scenario:
1. Employee check-in
2.Product flash sale online
3.Online sign-up for school programs/courses

Bussiness solution&Technical solution
1.Analyze the business scenarios to identify operations with hig QPS and TPS. List all relevant database tables involved along with their data charateristics.
2.Using a primary/replica setup is optimal. We can configure either a single primary with multiple replicas or multiple primaries with multiple replicas. This architecture is beneficial for distributing query operations during periods of high traffic, effectively managing increased loads.
3.Monitor Redis for hotkeys regularly and trigger alerts when detected
4.Design a multi-level caching architecture during application development to reduce the likelihood of Redis hotkeys forming. For example: in-memory session cache (e.g., Spring Session), local cache, and Redis cache.
5.Avoid using similar keys in the application. Having a large number of similar keys can increase the probability of hash collisions and reduce the efficiency of batch operations, as each key must be processed individually. In particular, when key sizes vary, it may lead to non-contiguous memory allocation and cause memory fragmentation.
6.Preload hot data into Redis before the application starts to prevent cache avalanche caused by high concurrent requests or cache misses.