Observability Driven Development

on YouTube ↗53m

" summary

Argues for testing in production and building observability to run real experiments: deploy A/B experiment with 50 blue variants (Google) and measure clicks, shadow new logic with the open‑source scientist pattern, and use feature-flag ring rollout to enable/disable risky releases. Recommends logs → metrics → distributed traces (Zipkin) to answer why a request slowed, defines concrete SLOs (example: 99.9% availability over 30 days) and ties observability metrics to deployment gates and product decisions.

AI-written summary · about

" chapters (9)

1. 0:00Start
2. 11:25WHAT IS KEY TO TESTING ON PROD?
3. 22:17TIME HAS COME
4. 23:29DEVOPS CYCLE
5. 27:53STRUCTURED LOGGING
6. 32:31METRICS
7. 34:00DISTRIBUTED TRACING
8. 39:47EXPERIMENT IN PRODUCTION
9. 52:43TONIGHT

" related talks

• 37mObservability-Driven DevelopmentCharity Majors · GopherCon Europe - Iceland 2018
• 44mObservability: Superpowers for DevelopersChristine Yen · Strange Loop 2019
• 31mObservability for Developers: How to Get ThereChristine Yen · DevOpsDays Seattle 2019
• 13mObservability that mattersTom Oketch · LeadDev Austin 2019
• 53mDeliver Results, Not Just Releases: Control & Observability in CDDave Karow · GOTO Chicago 2019
• 29mObservability for Developers: How to Get From Here to ThereChristine Yen · DevOpsDays NYC 2020
• 37mBetter Reliability Through Observability and ExperimentationJulie Gunderson, Kerim Satirli · KubeCon + CloudNativeCon Europe 2022
• 41mA peek into observability from testers lensParveen Khan · Appium Conference 2021