You Sell Caches? No, We Just Show Them
summary
Dodo Pizza faced high-frequency API requests with large response bodies (e.g., multi-MB mobile menus) that overloaded servers as user traffic grew, so the team implemented multiple caching approaches: in-memory and distributed caches, application-side response caching, and nginx-level caching. The talk describes the specific caching paths they tried, problems encountered for each, how they eventually made the most frequent API request operate without any cache, and their plans for future improvements.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:022019 performance issues
- 7:37System architecture evolution
- 12:12Caching menu problems
- 19:13Site attacks and issues
- 25:34Mitigating site attacks
- 32:55Tracking data problem
- 38:10Path Info service solution
- 42:44Data Catalog and DMS
- 51:19Two-year summary
- 54:26Conclusion and Q&A
related talks
35mCaching everywhere
24mCache for Commerce Performance and Optimization for High Volume Sales
43mCache Me Outside - Caching Methodologies & Architectures
21mHigh Performance APIs through Caching
25mCaching strategies on https://dev.to
47mCaching strategies on https://dev.to
24mCaching Strategy for High Volume Personalized Content Requests
42mCaching for Cash