Stage 1 of 10 · 6 lessons · ~3 weeks

Reliability basics: SLIs, SLOs and error budgets

Learn to describe reliability in numbers users would recognise: what to measure, what target to set, how much failure that allows, and what the team does when the budget runs out.

Start: Why 100 per cent is the wrong target

Lessons· one lesson for each topic, in this order

Do this

Pick one real service, even a side project, and write down who its users are.
Write two SLIs as good events over valid events and say where each is measured.
Set an SLO and a 30 day window for each, based on real data where you have it.
Convert each SLO to an error budget in minutes and in failed requests.
Write a half-page error budget policy and show it to someone who is not an engineer.

Start with· follow one of these

BookFree
Google Site Reliability Engineering Book
Google · Advanced
Free online: read the chapters on embracing risk and service level objectives first
BookFree
The Site Reliability Workbook
Google (O'Reilly) · Intermediate
Free online: the Implementing SLOs chapter walks through a worked example and an error budget policy

Also worth your time

High Availability in System Design: 15 Strategies for Always-On Systems (opens in a new tab)Read
Redundancy, failover and the nines behind availability targets, in one read
Implementing Service Level ObjectivesGo deeperPaid
The book on SLOs in practice: choosing SLIs, targets, budgets and getting an organisation to use them

Before moving on: A one-page SLO document for a service you run or have built: two SLIs written as good over valid events, an SLO and window for each, the error budget in minutes and requests, and an error budget policy.

Tauseef's advice

This stage is short in hours but it shapes everything after it. Every alert, dashboard and postmortem in this roadmap should point back to an SLO. If you have not deployed a service to Kubernetes with basic monitoring yet, do the DevOps and cloud roadmap first, then come back.