The whole category in one page.
ADOPT IN THIS ORDER
Version control with secrets excluded, tested backups, a documented deployment with a way to revert, then external monitoring that alerts someone.
Automated deployment without tested backups is faster destruction.
WRITE THE MANUAL PROCESS DOWN BEFORE AUTOMATING IT
That reveals every step nobody remembered. An automated wrong process is wrong faster.
SMALL FREQUENT RELEASES ARE SAFER
When a large release fails you cannot tell which change caused it, and a process used once a quarter is frightening rather than routine.
THE FASTEST ROUTE TO ANY CAUSE
What changed. Most incidents follow a change, which is why recording deployments matters.
If an incident followed a deployment, revert first and investigate with service restored.
THREE THINGS THAT FAIL SILENTLY FOR WEEKS
Backups, scheduled tasks, and certificate renewal. Monitor all three explicitly, and monitor business outcomes too — every technical check can pass while no orders have been placed.
NEVER MAKE A MANUAL CHANGE TO SOMETHING DESCRIBED IN CODE
And never broaden permissions to clear an error.
FIX THE SYSTEM, NOT THE PERSON
A system that depends on nobody making a mistake will fail.
WHAT MOST SMALL BUSINESSES NEVER NEED
Containers, orchestration, and elaborate pipelines. Most applications never outgrow one well-configured server.
