This site showcases build optimization experiments conducted using Telltale, an automated framework for running and analyzing Gradle builds. Here’s what you need to know:
How It Works
Experiment Execution: Telltale automates:
- Setting up experimental environments
- Running different build variants
- Collecting performance metrics
- Visualizing results
Data Collection:
- All builds are tracked in Develocity
- Build Experiment Results tool aggregates data
- Generates comprehensive CSV and HTML reports
AI-Powered Analysis:
- OpenAI analyzes experiment results
- Provides detailed summaries and insights
- Helps identify performance patterns and improvements
Publishing Process
When Do We Publish?
Experiments are published when:
- They are triggered by main workflows
- The
extra_report_args.deploy_resultsis set to'true' - Results are validated and ready for sharing
What’s in Each Post?
Each experiment post includes:
- Detailed performance comparisons
- CPU and memory usage analysis
- Build time metrics
- Task execution statistics
- Gradle User Home environment tags (e.g., #dependencies cache)
Technology Stack
- Develocity: Build performance monitoring
- setup-gradle action: Seeding Gradle User home for experiments
- Hugo: Static site generation
- Hugo PaperMod: Theme and styling
- Build Experiment Results: Data aggregation and analysis
Contribute Your Experiments
Have an open-source project? We welcome new experiments!
Requirements:
- Open source project
- Compatible with GitHub Actions free runners
- Interesting build configuration scenarios to test
Interested? Get in touch to start experimenting with your builds!
