SRE (Site Reliability Engineering)
SRE is a discipline that applies software-engineering practices to operations — using error budgets, SLOs, and blameless postmortems to make reliability a measurable, engineered property of a system rather than a matter of heroics during an incident.
SRE was originally coined at Google to describe treating operations as a software problem: instead of a separate ops team reacting to incidents, engineers define measurable reliability targets (SLOs) and use the gap between “100% perfect” and the target — the error budget — to decide how much risk the team can afford to take on new releases.
If a service is comfortably within its error budget, the team can ship faster and take more risk. If the error budget is nearly spent, the team slows down and prioritizes stability work over new features — a data-driven trade-off instead of a political one.
A common misconception is that SRE means “the on-call team.” On-call rotation is one piece of it, but the actual discipline is the measurement and decision-making framework around reliability — SLOs, error budgets, blameless postmortems that fix root causes instead of assigning blame, and toil reduction (automating repetitive operational work instead of accepting it as permanent).
Related terms