Performance for Conversion! Use Distributed Tracing to Pinpoint Bottlenecks

on YouTube ↗39m

" summary

Explains why Firebase Performance cannot reliably find root causes and presents using OpenTelemetry with a lightweight Kotlin SDK plus a collector pipeline (receiver → processor → exporter) to connect app spans to backend traces. Shows concrete measurements such as Time to First Frame and full render, splitting network timing into DNS/TCP/SSL phases, and a real example where a 3s full render was traced to a 500ms DNS delay; also discusses head vs tail sampling limits, inability to set trace start time in Firebase, and an app to distinguish user vs system launches.

AI-written summary · about

" chapters (11)

1. 0:00Introduction
2. 2:10Performance for conversion
3. 7:50Mobile app metrics
4. 10:44Firebase limits
5. 16:42Distributed tracing
6. 23:04OpenTelemetry tracing
7. 28:53Startup time measurement
8. 31:01Sending telemetry data
9. 33:51OpenTelemetry Collector setup
10. 34:59Results and benefits
11. 37:50Future work

" related talks

• 32mChasing performance issues methodicallyAmanjeet Singh · droidcon New York 2022
• 43mPerformance in a Kotlin WorldSean McQuillan · droidcon SF 2019
• 15mFix performance pitfalls and level up your app with FirebaseKim Faughnan, Nitin Kaushik · Google I/O 2021
• 45mPerformance in a Kotlin worldSean McQuillan · Chicago Roboto 2020
• 31mBuilding Consistently Fast Flutter AppsAbdullah Deshmukh · Fluttercon India 2025
• 39mDeep-Diving Android Performance with PerfettoAkniyet Arysbayev · droidcon Berlin 2025
• 20mOptimize app health with Firebase Performance Monitoring and CrashlyticsDan Chuparkoff, Jenny Cong, Michael Doyle · Google I/O 2022
• 16mHow to monitor errors and performance in Cloud Functions with Open TelemetryDaniel Lee · Firebase Summit 2022