Catch Me If You Cache: Cache Patterns and Pitfalls in Production
summary
The talk examines caching in production: what a cache actually is, common caching patterns (and their trade-offs), classic anti-patterns that cause operational nightmares, and a decision tree for selecting the right caching strategy. It covers practical issues like eventual consistency, stale data, distributed side effects and hard-to-reproduce bugs, and notes similar challenges arising with LLM-related approaches such as semantic caching.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 7:44Cache definition purpose
- 13:17Three cache trade-offs
- 16:26Critical cache questions
- 19:36Cache placement strategies
- 21:00Cache update strategies
- 25:37Invalidation vs eviction
- 26:41Thundering herd problem
- 28:26Cache antipatterns
- 30:52Cache decision grid
- 33:22Best practices monitoring
- 35:29Caching for LLMs
- 40:44Semantic caching for LLMs
- 42:31Conclusion
related talks
43mWhere Is My Cache? Architectural Patterns for Caching ...
46mA Hitchhiker’s Guide to Caching Patterns
09mCache 101: Understand key strategies before applying them
27mCaching for Cash
42mCaching for Cash
43mCache Me Outside - Caching Methodologies & Architectures
47mWeb Caching: Strategies, Pitfalls and Modern Tools
01h 10mCaching with Spring: Advanced Topics and Best Practices