Practices that prevent incidents.
WHAT TO DO BEFORE EVERY DESTRUCTIVE COMMAND
Read the path aloud, and confirm which machine you are on.
WHAT TO DO BEFORE EDITING CONFIGURATION
Copy it to a dated backup.
WHAT TO DO BEFORE RESTARTING A SERVICE
Test its configuration.
WHAT TO DO BEFORE ANY LONG TASK
Start a persistent session.
WHAT TO DO BEFORE CHANGING FIREWALL RULES
Arrange a way back.
WHAT TO DO AFTER ANY CHANGE
Verify the thing you changed, and one thing you did not.
WHY THAT SECOND PART
Changes have consequences beyond their intent.
WHAT TO RECORD
What you did, when, and why.
WHERE
Somewhere another person could find it.
WHY IT MATTERS
You will not remember in six months, and someone else may need to.
WHAT TO AUTOMATE ONCE YOU HAVE DONE IT THREE TIMES
The task.
WHAT TO MONITOR
Anything whose failure you would want to know about before a customer does.
WHAT TO REVIEW MONTHLY
Disk space Backup success Certificate expiry Available updates
WHAT TO PRACTISE WHILE NOTHING IS WRONG
Restoring a backup.
WHY
Under pressure is the worst time to learn a procedure.
WHAT TO ACCEPT
That care is faster than recovery.