• Sources: primary
  • Summary: For the 10th anniversary, Google published a completely rewritten second edition. It covers multi-tiered SLOs with burn-rate alerting, OpenTelemetry and tracing, reliability of GPU and TPU clusters and training pipelines, and generative AI for incident triage and postmortems. It is available in print and on O'Reilly.
  • Why it matters: The reference text most SRE practices cite now reflects current observability and AI infrastructure practice.