Automation design: when to automate and how to fail safely
For systems administrators, DevOps engineers and MSP technicians who write scripts, RMM automations or runbook automation that act on production systems, and for the leads who approve them. You should be able to read a simple script; no particular language is assumed. You'll learn to recognise and measure toil, decide whether a task should be eliminated, rejected, simplified or automated, build in the safeguards that stop automation failing at scale (validated inputs, scope and rate limits, idempotent steps, a stop control), decide where people stay in the loop, and write a design review that a colleague can approve. Grounded in Google's SRE book and workbook and the AWS Well-Architected operational excellence principles; your change process and access policy decide what automation may do and who signs it off. Ends with a supervisor-graded design review case.
- Level
- Intermediate
- Length
- About 90 minutes
- Contents
- 5 lessons · final exam
- Status
- Published · updated 10 Oct 2026
Skills you'll practise
- Identify toil in your team's work using its characteristics, and measure it before deciding what to automate
- Decide whether to eliminate, reject, simplify, partly automate or fully automate a task, weighing cost, consistency and risk
- Design automation that fails safely, with validated inputs, scope and rate limits, idempotent steps and a stop control
- Choose where a human stays in the loop, and keep operators able to understand and run the process when automation fails
- Write an automation design review that records triggers, guardrails, failure behaviour, ownership and monitoring
Course outline
- 1.Identify toil and measure it before you automateLesson · 14 min
- 2.Decide what to do about a task: eliminate, reject, simplify or automateLesson · 16 min
- 3.Design automation that fails safelyLesson · 18 min
- 4.Keep humans in the loop and able to take overLesson · 14 min
- 5.Write an automation design reviewLesson · 14 min
- 6.Automation design: when to automate and how to fail safely: knowledge checkKnowledge check · 14 questions
- 7.Automation design: when to automate and how to fail safely: practical exerciseKnowledge check · 1 question
- 8.Final exam10 questions · passing it completes the course, so people who already know the material can test out
Sources it draws on
The lessons and questions are written from these references, so learners can go back to the original.
- Google SRE book, Chapter 7: The Evolution of Automation at Google (value of automation, automation classes, Prodtest and idempotent fixes, Diskerase)
- Google SRE book, Chapter 5: Eliminating Toil (definition and characteristics of toil)
- Google SRE Workbook, Chapter 6: Eliminating Toil (measuring toil, taxonomy, toil management strategies, assessing risk within automation)
- Google SRE Workbook, Chapter 10: Postmortem Culture (the satellite Diskerase incident: root cause, missing safeguards and action items)
- AWS Well-Architected Framework: Operational excellence design principles (safely automate with guardrails; frequent, small, reversible changes)
See it with your own jobs and topics
Tell us about your team and we'll walk you through setup, from choosing jobs to your first skills check.