Explain like I'm five · SRE

What is a Runbook?

A runbook, told as a story about the mechanic's notebook that spells out 'do this, then this' for every common failure.

how did we fix this last time...the senior is on vacationthe same failure hits again, and nobody remembers how it was fixed last time

The same failure hits again, and nobody remembers how it was fixed.So they start from scratch, confused.

(it's only in my head)I don't know itif it's only in someone's head, nobody knows it on their day off

If it's only in someone's head, nobody knows it on their day off.The rookie is left stuck.

THE NOTEBOOKa ride won't startthe line won't shrinkthe screen goes darkfor each common failure, write down 'do this, then this' in a notebookwhoever's on duty can just follow the notebook

For each common failure, write down 'do this, then this' in a notebook.Whoever's on duty can just follow the notebook.

Write down the common failures firstrare ones can wait
Step by step, picture-book simpleone clear line at a time
Simple enough for a rookieno extra explaining needed
Add what you just learnedincluding what came out of a postmortem
WHEN A RIDE WON'T START1. check the red button first2. check the warehouse stock3. if not, page the seniorstep by step, simple as a picture book — even a rookie can follow it

Open the notebook, and it's all numbered in order.Follow the order, and even a rookie can do it.

OLD NOTEBOOK1. the old way(doesn't work anymore)left unfixed too long, it stays wrongUPDATED NOTEBOOK+ what we just learnedupdate the notebook every time something changes

Left unfixed too long, the notebook stays wrong.Every fix needs to update the notebook too.

In one breath

Runbook = a notebook spelling out the steps for each common failure, so whoever's on duty can fix it just by following along.

A document that spells out step-by-step responses to a recurring failure, also called a playbook. It can become the script for automation, helps onboard newcomers, and stays useful only if lessons from postmortems keep getting added.

When grown-ups say it

Runbook (playbook)
The notebook with steps for each failure. Also called a playbook.
Step-by-step procedure
The order written out as 1, 2, 3. Just follow it in order.
Relationship to automation
How the notebook eventually becomes an automation script. Repeating the same steps is a sign to automate them.
Rookie onboarding
How a newcomer learns from the notebook. They can follow along without extra explaining.
Keeping it current
Not leaving stale content as-is. Left unfixed, it becomes a 'stale runbook'.
Checklist
A list that makes sure nothing gets skipped. Often sits right inside the runbook.
Link to the postmortem
Adding what was learned from an incident into the notebook. → the record everyone writes without blame
Toil
The signal that shows up when the notebook keeps growing. The same repetitive chore, every day. → the chore that repeats every day
Next story
The Record Everyone Writes Without Blame →
Want another term as a picture book? Request →