The state of SRE in 2022
summary
Argues that SRE has no single definition but is clearly rising in interest; cites concrete tool uptake: roughly five people report using eBPF, a majority run Kubernetes, and observability stacks like Prometheus and Grafana plus Terraform are mainstream. Recommends embedding SRE early (design/release), make metrics part of the definition of done, and pick an operating model by scale—embedded, horizontal, or federated—while optimizing for lower context‑switch costs and avoiding a “kitchen sink” SRE that becomes perpetual firefighting.
AI-written summary · about
chapters (10)
- 0:00Introduction and SRE overview
- 3:12Defining SRE challenges
- 7:45SRE role ambiguity and labor division
- 9:10SRE growth and tooling revolution
- 12:19CNCF ecosystem and open source
- 14:24Standardization and future of SRE
- 17:33Q&A: SRE model preferences
- 21:47Q&A: Developer vs SRE handoff
- 26:45Q&A: Convincing executives
- 28:33Q&A: Federated vs central SRE
