It’s not recklessness, it’s science
Ben Eggers has lived through a lot of outages, and he thinks you should test in prod.
The sad truth is that testing is (kind of) optional. Deploying to production is not. Every time you ship new code, you’re testing in production anyway – so, why not do this with intent?
Intent doesn’t mean unplugging your database and hoping for the best, but developing crisp hypotheses, good metrics, and recovery plans – and then unplugging your database. Ideally, you also want someone else to be on-call.
These are the details that make the difference between reckless and responsible.
The reason to do this is that there’s really no place like prod – and to prove the point, Ben wrote a little game called “Guess the Impact.” You should watch the talk and play the game, and if you can do better than the 200 reliability geeks at BugBash, we’d really like to talk to you.
