The Invisible Door: Reliability Gaps in the Front End
summary
Front end reliability focuses on crashes, TTI, jank and ANRs, not just CPU, and often requires high‑cardinality metrics and heavy sampling. A pilot team replaces 20 engineers juggling issues with three dedicated roles, confronting 20,000 device models and Google Play Store core‑vitals SLAs. Off‑the‑shelf metrics cost estimates hit $120M/month for distinct metrics, so the engineered solution uses a custom tracer that emits OTLP JSON to the otel Collector and Honeycomb Refinery, plus OTA, feature‑flag kill switches and per‑trace heavy head/tail sampling.
AI-written summary · about
chapters (9)
- 0:00Front end reliability intro
- 3:14Why SREs care about front end
- 9:11Ownership gaps in client reliability
- 15:36Front end reliability metrics
- 25:35Front end release cycles
- 33:03Front end observability costs
- 36:58Tracing front end systems
- 41:56Custom tracing solutions
- 42:39Collaboration and future
related talks
39mThe Frontiers of Reliability Engineering
32mObservability for Frontend Engineers
38mThe 15 Minute Model for Optimising Device Experiences
30mAvoiding software fails: few metrics to improve application reliability
22mFear of the computer
38mAre Device Response Times a Neglected Risk of IoT?
26mBeyond the Hype: Facing Micro Frontend Challenges in Large Scale
11mAre They Really Using It? Monitoring Digital Experience To Determine Feature Effect