Company
Date Published
Author
Dan Holloran, Jemiah Sius
Word count
2170
Language
English
Hacker News points
None

Summary

At New Relic, programmatically tracked service level indicators (SLIs) are foundational to site reliability engineering practice, reducing manual tracking of performance and incident data. Programmatic SLIs have three key characteristics: current, automated, and useful. They help teams develop and create programmatic SLIs by identifying system capabilities, articulating plain-language definitions, defining one or more SLIs, measuring the system, defining SLOs, iterating, and refining over time. New Relic One's service level management functionality can automatically set up SLIs and SLOs, providing a one-click setup for establishing a baseline for reliability metrics. To track programmatic SLIs, teams identify existing instrumentation, deploy new instrumentation, gather metrics, create alert conditions, and create NRQL queries and dashboards to reveal when services miss their indicators. Regularly revisiting SLIs is necessary as systems evolve, and it's recommended to revisit them quarterly or whenever changes occur.