Gradle Build: Measuring What Matters Most
summary
A presentation about measuring and analysing Gradle build performance for Android using a custom Gradle plugin developed at Tinkoff. It describes which build metrics the plugin collects, what insights those metrics revealed about build-time regressions and impacts of changes, how the plugin is implemented and what Gradle-provided information can be extracted, and offers the plugin for open-source use in other projects.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 1:48Why measure Gradle builds
- 5:23Custom plugin reasons
- 8:20Writing metrics plugin
- 10:10Gradle plugin types
- 11:54Build phases and data sources
- 13:19Gradle public vs internal API
- 15:50Build operation hierarchy
- 17:13Build operation count
- 18:39Internal API limits and advice
- 19:21Gradle config cache issues
- 21:10Metrics problems and fixes
- 22:18Plugin implementation and setup
- 23:40Use cases and tasks
- 24:46App size and artifact metrics
- 26:33CI vs local build times
- 27:37Metrics trend analysis
- 28:43Build environment impact
- 30:55Android Studio sync optimization
- 32:01Configuration on Demand
- 33:48Machine impact on builds
- 34:50Build time factors
- 35:13Reusing metrics solution
- 36:40Data processing and visualization
- 37:47Code resources
- 38:10Conclusion
related talks
28mMaking the most of your Gradle Builds
36mImproving Gradle Builds
33mIMPROVING ANDROID BUILD PERFORMANCE
30mImproving Android Build performance
43mThe Secrets of the Build Scan Plugin and the internals of Gradle
37mThe Build Shrugged
43mExperiments for your Android builds driven by Gradle Profiler
31mImproving Android Build performance