Great Caching Disasters!
summary
Production outages trace to caching: a poll-triggered automated cache clear repeatedly drops a site serving 2,000 req/s on 12 servers, and a missed notification leaves a deleted article cached under a three-day cache so half the site shows “sorry” pages for days. HTML fragments with server-side includes causing per-fragment DB calls break under live Q&A load. Moving to an automated, simple one-minute HTML fragment cache fixes most problems; long or manual caches and processing-in-cache are the real hazards.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:33Caching Disasters
- 2:00Two Important Lessons
- 2:24Build a Clear Cache
- 2:45Long Cash Times
- 3:05Three Day Cash
- 7:20Woolly Rat
- 8:45Conclusion
- 9:10Lessons
related talks
27mFAIL: The Best Ways To Bring Down Your Website
25m"The website's down!" Stories and lessons on keeping your website up.
36mMake It Fast
25mCaching strategies on https://dev.to
48mPerformance Mistakes
24mCaching Strategy for High Volume Personalized Content Requests
43mCache Me Outside - Caching Methodologies & Architectures
39mMaking it faster