Dealing with Verification Data Overload in OpenJDK Projects

Explore strategies for managing overwhelming verification data in open-source Java projects during this conference talk. Learn about data visualization, filtering, and categorization techniques used by IBM engineers to process massive amounts of test logs and console output generated daily in AdoptOpenJDK and Eclipse OpenJ9 projects. Discover how deep learning technologies are being applied to refine raw data, efficiently display results, and support complex continuous delivery pipelines. Gain insights into the AdoptOpenJDK Quality Assurance framework, test grouping methodologies, and the Test Result Summary Service (TRSS) for personalized dashboards and pipeline monitoring. Understand the challenges of dealing with multiple versions, platforms, and implementations in Java testing, and learn how to effectively narrow down problems and improve farm monitoring.

Intro
The Scope
Degrees of Freedom
Gather Great Tests
AdoptOpenJDK Quality Assurance (AQA)
AQA Manifesto
Test Framework (TestKitGen)
Grouping & Granularity
AdoptOpenJDK CI Pipeline
Jenkins scripts for testing
TAP & JUnit Plugin
Archive data
Test Result Summary Service (TRSS)
TRSS Overview
TRSS: Personalized Dashboard
TRSS: Monitor Jenkins Pipeline Builds
TRSS: Test Builds Result
TRSS: Tests Result
TRSS: Search test
TRSS: Perf Dashboard
Let Us Count the Ways
What is Deep Learning?
Initial DL Experiments
Model Building
Plans Forward
Connect & Get Involved