Blog
- Why we still write runbooks
Every incident we have handled faster than expected had one thing in common: someone had written down what to do before it happened.
- Backups you have never restored do not exist
We start every new engagement by asking when the last restore test happened. The honest answer is often never.
- Keeping deployments boring
A good deployment is one nobody talks about afterwards. We aim for small, frequent releases that can be rolled back with a single command.
- Choosing a database for a small product
Teams often ask us which database they should start with. Our answer has not changed in years: start with PostgreSQL unless you have a concrete reason not to.