Brian Louis Ramirez at performance.sync()
summary
A project monitors customer sites for RUM incidents and regressions, proving that specific incidents beat broad topic scans: an actual mobile Safari crash is found by comparing AB tests and by correlating Hex + SQL + Python queries with service‑worker and custom timings. Rollouts start at 10% so control groups exist; 100% rollouts lose that signal. An hourly aggregate with z‑score anomaly detection pushes Slack alerts and dashboards, iterates via rapid prototypes and works well for catching regressions early.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:29Performance haystacks
- 5:08SpeedKit setup
- 7:18Mobile Safari incident
- 11:18Iterative development
- 14:12Anomaly detection system
- 15:14Key takeaways and closing thoughts
related talks
18mProduct Reliability in
02mMobile Real User Monitoring
20mOptimize app health with Firebase Performance Monitoring and Crashlytics
34mWhy Google’s CrUX Results Are Not Reproducible With Your Real-User Monitoring
14mBeyond the Basics: Performance Monitoring and User Experience for Mobile App Growth
14mWeb Apps, through the Lens of SRE (Site Reliability Engineering)
44mEfficient Android app monitoring
16mHome Brewing R.U.M - Analyzing application performance with real user monitoring