Testing in prod has gotten a bad rap. It’s both inevitable - you can’t know everything before you ship - and desirable. In modern complex systems, failure is a constant and the only guiding principle is that “users should never notice”. So how do you test safely in prod, and how should you allocate your scarce engineering cycles between prod and staging? Charity is the cofounder of honeycomb.io and co-author of database reliability engineering. her mission in life is to …