Troubleshooting Gradle and Apache Maven Stefan Wolf Alex Semin Oleg Nenashev with Develocity

No content

No content

Hello, World! Stefan Wolf Build Scans Team / Jenkins plugins maintainer 4 Alex Semin Configuration Team / BT User Success Group Oleg Nenashev DevRel & Education Team / Dev Tools community builder

Async Q&A: Gradle Community Slack, Testcontainers and WireMock Slack speakerdeck.com/onenashev

Today, 2nd of December 6 ● The Presentation and Demos - Live ● Q&A after the presentation - Live ● Any meetings/discussions - in Munich ○ Share your Gradle Build Tool experiences with Alex ○ Ask Stefan about Build Scans and Develocity ○ Any questions about Gradle! ● Later: Online meeting with other offices?

Talk Outline ● Some Mythbusters ● Intro to Gradle Builds Troubleshooting ● Show & Tell for Gradle and Apache Maven 7 The information contained in this presentation and the roadmap (and the "Roadmap") is intended to outline our general product direction, which is subject to change at any time. The content provided in this Roadmap is provided for informational purposes only and is not a commitment, promise or legal obligation to deliver any material, code, or functionality and should not be relied upon in making purchasing or other decisions. The development, release and timing of any features or functionality described in this Roadmap remains at the sole discretion of Gradle, Inc. Product capabilities, timeframes and features are subject to change and should not be viewed as commitments.

The Elephant in the Room? Amadeus loves Apache Maven, … and we too!

No content

Disclaimer 10 Gradle != Gradle Build Tool in this presentation It’s not a “Gradle vs Maven” talk Gradle is about both Maven and Gradle Build tool

1 1 And we mean it Valentine Day’s post

+

13 See ge.apache.org helps Used in Beam, Groovy, Kafka, Lucent, Pulsar, Solr, etc. ● 37 onboarded projects ● 31 days per week - savings from the build cache ● 116 days per week - savings from all caching ● 30k builds per week ● 503d of total build time per week

What is Gradle, really?

15 Developer Productivity Engineering (DPE)

16

No content

What do Gradle users say? (*) * based on user feedback, not Android specific

Developer Productivity Engineering (DPE) gradle.com/ developer-productivity-engineering

20 (f.k.a. Gradle Enterprise)

TL;DR: * we do not actually suggest moving

22 https://gradle.com/develocity/

DevelocityTM Build Scan® “X-Ray” for Gradle, Maven, Bazel, & sbt (CI & Local) Platform Flaky Test Management Build Scan & Comparison Performance & Failure Reporting Vulnerability & Compliance Reporting Out of the Box & Custom Reporting & Visualization Test Failure Analytics Benefits Test Distribution AI-Powered Predictive Test Selection Build Cache | Developer Productivity Platform Observability & Traceability Build & Test Acceleration Faster Troubleshooting

We gave 20,000 developers back two hours per day by cutting Maven build and test times by over 60%. Top 3 Global Investment Bank

For many FOSS projects: we sponsor Develocity licenses / SaaS hosting help with optimization and DPE Maven, Bazel and sbt users welcome! gradle/develocity-oss-projects gradle/develocity-oss-projects

Troubleshooting Gradle

No content

Source: https://www.brainyquote.com/quotes/naoto_kan_510414

So, we… 1. Discover the issue 2. Analyze the issue 3. Apply the fix 4. Verify the issue is gone 29

So, we… 1. Discover the issue 2. Analyze the issue 3. Apply the fix 4. Verify the issue is gone 30 Observability

31

32 Gradle Core Features ● Gradle includes many troubleshooting capabilities ● We work on improving reporting and stack traces ● NEW: Problems API and reporting

Problems HTML Summary Report in Gradle 8.11

For performance profiling 34

35

Develocity | Observability is the Foundation Observe Accelerate Troubleshoot 36

37

○ Build Scan and Troubleshooting ○ Local and Distributed build caching ○ Flaky test detection ○ Local build observability ○ // And more! 38 for Gradle and Apache Maven Troubleshooting

Flaky Tests 39

a. Try it again b. Re-run it c. Re-run it again d. Ignore it and approve PR e. All of the above The test is flaky. What do you do now? 40

No content

Flaky test handling ⬢ Detection ○ Test retries ○ Develocity ⬢ Flaky tests don’t fail the build ⬢ Fixing ○ New very flaky test on the spot ○ Regular flaky fix-it days 42

43 https://ge.apache.org/scans/tests?search.buildToolType=maven&search.relativeStartTime=P28D& search.timeZoneId=Europe%2FBerlin&tests.sortField=FLAKY

Underutilized CPU resources 44

45 https://gradle.grafana.net/d/gradle-cpu-usage/gradle-cpu-usage?o rgId=1&from=now-7d&to=now&timezone=browser

46 https://ge.gradle.org/s/3spxtqmeeioze/timeline

Build Caching 47

:compileJava :compileTestJava inputs actions outputs Gradle Tasks / Maven Goals 48

:compileJava :compileTestJava 💻 Gradle Tasks / Maven Goals 49

:compileJava :compileTestJava 💻 UP-TO-DATE FROM-CACHE Build Cache Build Cache / Maven Goals 50

51

Demo Time! 52

Live Demo: What can we show? ● Analyzing slow smoke tests ● Surfacing flaky tests in the Spring Boot or Armeria ● Analyzing Build and Configuration cache speed ● Discovering differences on CI generated jars causing integration tests to rerun - Micronaut, / Maven ● /// and More! 53 gradle/develocity-oss-projects

If you watch the slides offline … 54

Build Time Reduced by 50% & Test Time by 60% Case Studies | Netflix Total test time reduced 60% Even Faster Total test time reduced 60% Build Time Dropped 50%!

110+ Days in CI and Local Test Savings with Predictive Test Selection in 30 Day Period L o c a l CI Serial testing time that could have been avoided (purple) vs unavoidable (gray) in a single day 33.3 days of serial testing time that could have been avoided if PTS was enabled 1.1M tests (42% of all tests) that could be avoided if PTS was used Local Builds CI Builds

14% of builds that have had at least 1 test fail 9% of builds that have had at least 1 flaky test (same test fails then passes in the same build) Sorting by Flaky column will show the test classes with the most flaky tests Flaky and Failed Tests Identification and Prioritization

Test Times Before and After Develocity Test Distribution | 2 Projects Case Studies | Netflix After Test Distribution Before Test Distribution Impact ○ (2) Projects ○ 33 Tasks ○ Before: ○ 1 hour 2 min ○ After: ○ 4 min 59 secs ○ Savings: ○ 57 min 1 sec

The End? 59

What can you do? ● Ask your DevEx team about Develocity ○ Ongoing deploy/enablement ● Share your feedback with us! ● Learn Gradle Build Tool, Apache Maven and Develocity on DPE University 60

dpeuniversity.gradle.com DPE University Take a dive into Gradle! 61

6 Gradle courses for all levels of experience DPE University dpeuniversity.gradle.com 62

dpeuniversity.gradle.com DPE University Maven & Develocity 63

References ● Gradle Build Tool roadmap: https://roadmap.gradle.org/ ● Develocity for Gradle Build Tool: gradle.com/gradle-enterprise-solutions/gradle-build-tool/ ● Develocity fort Apache Maven: https://gradle.com/gradle-enterprise-solutions/maven-build-tool/ ● Public Develocity instances for FOSS projects: https://gradle.com/customers/oss-projects/ 64

Gradle Community Slack, #developer-productivity 65 Open Questions?