a pint sized introduction to SLO
summary
Argues that reliability needs concrete SLOs tied to business metrics not vague dashboards, claiming internal SLO targets should be stricter than customer SLAs and using five nines (~6 minutes/year) as an example of a risky goal. Recommends instrumenting applications for a condensed set of metrics that answer whether customers are shopping and why, using the three pillars: metrics logs traces to diagnose issues, and monitoring burn rate to decide when to page versus ticket (backups versus login failures).
AI-written summary · about
related talks
35mAvoiding Goodhart's Law - Use SLO's as Tools Not Cudgels
28mLean Product Development Through SLOs
38mImplementing SRE practices: SLI/SLO deep dive
30mMetrics for good, not evil
31m9 Things You Should Do When Starting to Use SLOs
01h 06mPepperoni, Puppets, and Priorities: Demystifying SLAs, SLOs, and SLIs
22mLook Ma, No Metrics: A Quest Towards Extracting SLOs
09mHow to answer tricky questions the SRE way