SLX - An Extended SLO Framework to Expedite Incident Recovery

Explore an extended SLO framework designed to expedite incident recovery in this 31-minute conference talk from SREcon21. Dive into the journey of establishing Service Level Objectives (SLOs) for an infrastructure SRE team with a 99.999% availability target. Learn about the process of defining SLOs, understand the gaps between expectations and reality when working with development teams, and discover a unified SLO framework called SLX. Examine how SLX facilitates the management of hundreds of SLOs, incorporates statistical anomaly detection algorithms, and introduces new concepts like Service-Level-Factor (SLF) and Service-Level-Dependency (SLD). Gain insights into building SLO knowledge graphs across multiple infrastructure systems and explore the intent-driven SLX implementation inspired by Kubernetes design and GitOps paradigm. Follow the presentation through various topics including incident lifecycle, SLX definitions, automation, workflow, system design, dashboards, and anomaly detection, concluding with key takeaways for improving incident recovery processes.