Explore the intricacies of constructing accurate latency Service Level Objectives (SLOs) in this 31-minute conference talk from SREcon19 Americas. Delve into three distinct approaches for calculating latency SLOs correctly, and discover how histograms can be utilized to compute mathematically precise quantiles for setting SLOs. Learn why common metrics like median, average, 90th, and 99th percentile are often misapplied in monitoring systems, both open-source and commercial. Gain valuable insights from Fred Moyer of Circonus on implementing latency SLOs that truly reflect service performance and reliability.
Overview
Syllabus
SREcon19 Americas - Latency SLOs Done Right
Taught by
USENIX
Related Courses
-
Latency SLOs Done Right
-
Latency SLOs Done Right
-
Latency SLOs Done Right - Calculating and Setting Accurate Service Level Objectives
-
The Map Is Not the Territory - How SLOs Lead Us Astray, and What We Can Do about It
-
SLOs for Data-Intensive Services
-
Methods to Achieve High SLOs on a Large Scale Kubernetes Cluster
