FAIL: The Best Ways To Bring Down Your Website
summary
Concrete failure stories and fixes: sites hit sustained peaks of 38,000 requests/sec (from ~12,000) forcing rapid fixes and extra machines; WordPress defaults store images on local disk so scaling needs CDN or plugins; a 2.4MB hotlinked image from a tiny site brings down both host and publisher; careless polling change from 60s to 5s turns thousands of users into an accidental DoS; single slow TCP connections can exhaust Apache threads; build redundancy, test limits, and force-fail parts of production.
AI-written summary · about
chapters (9)
- 0:00Start
- 5:09Why things FAIL
- 8:26How much traffic did you expect?
- 9:38Plan for failure
- 12:43Scalability
- 13:41Scaling
- 15:05Web architecture #1
- 15:47Web architecture #3
- 16:55Web architecture #4
related talks
10mGreat Caching Disasters!
25m"The website's down!" Stories and lessons on keeping your website up.
30mAvoiding software fails: few metrics to improve application reliability
41m12 Factors of Pain and Suffering
53mSuccess at scale: Preparing Drupal 8 for large events
46mDrupal’s Scariest Things
33mFrontend Resilience
28mMy Favourite Errors