When a Deploy Breaks Production
Roll back first, debug later, and keep rollbacks safe with flags
What you'll learn
- Why rollback comes before debugging
- Database changes that keep rollbacks safe
- Feature flags and canary releases
- Automatic rollback and blameless postmortems
In this video, we'll look at what to do when a new deploy goes out and errors suddenly jump to 30%. We'll cover why you should roll back before debugging, how database changes can make rollbacks hard, and how feature flags let you turn off a broken feature without deploying again.
We'll also look at canary releases, automatic rollbacks, and blameless postmortems.
By the end, you'll know how to recover from a bad deploy and keep the next one from reaching all your users.
