How does site reliability engineering differ from traditional devops in a cloud environment?

All answers to this question.

The most common saying is that "class SRE implements interface DevOps." While <devops> focuses on breaking down silos between dev and ops to increase delivery speed, SRE focuses on using software engineering principles to solve operational problems. In a daily routine, an SRE spends 50% of their time on "toil" (manual ops tasks) and 50% on project work to automate that toil away. You should definitely study Service Level Objectives (SLOs) and Error Budgets. An SRE’s job is to define how much "unreliability" is acceptable before the team must stop shipping new features to focus on stability.

   Answered 2025-01-22 by Helen Roberts

Are you prepared to handle "on-call" shifts, as that is often a much larger part of the SRE life compared to a standard <devops> role?

   Answered 2025-01-24 by Walter White

• Walter, I am okay with on-call as long as there is an automated recovery process. In your experience, do SREs get more autonomy to fix the root cause of an outage compared to a regular engineer? I’m tired of just restarting servers; I want to actually fix the memory leaks or bad queries that are causing the crashes in the first place.

     Commented 2025-01-25 by Harold Ford

SRE is much more data-driven. You’ll spend a lot of time in Prometheus and Grafana dashboards compared to a pure <devops> automation role.

   Answered 2025-01-27 by Marie Johnson

• I agree with Marie. If you love metrics and fine-tuning system performance, SRE is the way to go. It’s a very rewarding path for those with a deep technical itch.

     Commented 2025-01-28 by Christine Evans