SRE

Site Reliability Engineering (SRE) is a discipline that uses software engineering to solve infrastructure and operations problems, with the goal of building scalable, reliable software systems. It was developed at Google in the early 2000s to bridge development and operations and has since spread to many organizations.

A core SRE concept is aligning reliability with product goals through quantitative targets: service level indicators (SLIs) measure performance or availability, service level objectives (SLOs) specify target values, and error budgets cap how much unreliability is acceptable. The idea of toil—manual, repetitive operational work—drives automation and programmatic tooling to reduce incidents that consume time.

Practices include monitoring and alerting, incident response, and blameless postmortems focused on learning. SREs perform capacity planning, disaster recovery, and release engineering to ensure safe deployments and scalable operations. The aim is to automate routine tasks and enable developers to ship features without compromising reliability.

In organizations, SREs are software engineers who collaborate with product teams to define SLOs and implement reliability improvements. A blameless culture and data-driven retrospectives are central. SRE is often presented as a concrete implementation of DevOps principles, emphasizing engineering discipline and measurable reliability rather than process alone.

Common metrics include uptime, P95 or P99 latency, and error rates. Adoption varies; some teams use platform or reliability engineering groups in place of traditional operations. Practices adapt to organization size, product domain, and technology stack.