The code that passes QA is not the code that survives production. This is the field manual for what happens after launch.
About the book
Michael Nygard's premise is blunt: every system will eventually be pushed past its design limits, and the only question is whether it degrades gracefully or collapses catastrophically. Production-ready software isn't just correct, it's resilient, observable, and keeps operating through partial failure without a human paging in at 3am. The book catalogs stability patterns — circuit breakers, bulkheads, timeouts — alongside capacity planning and decoupling deploy from release.
Best used for
—Adding circuit breakers, bulkheads, and timeouts before one integration point takes the whole system down
—Planning a zero-downtime deploy with release decoupled from deploy
—Load-testing past peak and running failure injection before the postmortem writes itself
“This remains one of the most important books software engineers can read.”
— Rod Hilton, Goodreads
Install this skill
Run this command in your terminal. The skill will be available in all your Claude Code projects.