Live Q&A Session #2
summary
Live Q A summarizes concrete reliability practices and outcomes not platitudes it shows using Auto Canary with automated enrollments and exponential adoption to reduce config-caused outages, using Canary spec doctor to skip unhealthy specs and auto-notify owners while planning automated fixes, and using annotations plus a monthly SLO violation review to turn spikes into diagnosable insights with automation via Dr Patterson The stack runs A B and short full‑production degradations for 3 billion users and reports site‑level degradations about once per year
AI-written summary · about
chapters (12)
- 0:00Event overview
- 0:47Defining reliability
- 3:59Convincing teams to annotate SLO violations
- 6:49Service degradation at massive scale
- 9:17Production testing essentials
- 10:48Service degradation frequency
- 12:13Safer config changes with Auto Canary
- 14:56Scaling Canary specs
- 16:19Balancing speed and safety
- 18:46Accountability in blameless SEV culture
- 21:10Recognizing SEV contributions
- 25:03Closing remarks
related talks
24mLive Q&A with Speakers – Laine Campbell
19mLive Q&A
27mLive Q&A with Speakers – Marina Fisher
18mProduct Reliability in
19mFireside Chat with Jason Yee
43mBest practices for continuous testing and observability
14mFrom Dashboard to Defense: Automating Resilience at
17mAI-Driven Prompts for Enhanced System Reliability