How Google SRE and developers work together
summary
Google’s SRE model scopes work as baseline/assisted/full engagements funded by developer orgs, with SRE product areas of ~50–300 engineers inside a multi‑thousand‑person org. Concrete practices include using SLOs and error budgets to trade reliability for velocity, a baseline “esre love” offering of 2–4 hours/week time‑boxed consulting, embedding one or two SREs for assisted projects, a minimum core team of two sites with ~6 SREs each for deep support, an automate yourself out of the job every 18 months goal, and “code yellow” all‑hands to fix systemic failures.
AI-written summary · about
chapters (12)
- 0:00Google SRE introduction
- 1:12Reliability vs feature velocity
- 3:38What is Google SRE
- 6:49SRE engagement model
- 10:47Deciding SRE work
- 12:36Measuring success
- 15:25Reliability in development
- 17:09Baseline SRE support
- 19:16Assisted SRE projects
- 21:01Full SRE support
- 23:32Handling overload
- 24:59Google SRE summary
related talks
40mHow Google SRE and developers work together
32mHow Google SRE and Developers Work Together
41mAn Introduction to SRE at Google (with DevOps Focus)
18mSRE Engagement Model Transition in Building and Expanding SRE Team
22mSRE for Enterprises
47mSRE Team Lifecycles
05mHow do Devops and SRE relate?
27mUnified Theory of SRE