Flaky Tests, Slow Builds, Overloaded CI? You're Guessing at the Problem
summary
Claims that observability with Gradle’s Dev-Velocity reveals concrete causes: 30–40% of CI time spent downloading dependencies, 90% of CI CPUs idle despite queueing, and a small set of tests causing most flakiness. The tool surfaces build counts, trends, flaky-test culprits, unknown dependencies, and supports LLM queries to ask “why did this build fail.” Proven fixes include adding caching to cut dependency time and fixing the few high-impact tests to restore trust and throughput.
AI-written summary · about
chapters (8)
- 0:00Performance tuning: Measure, don't guess
- 1:40Common developer experience pain points
- 2:49Rethinking the strategy: Beyond throwing hardware at problems
- 4:41Monitoring versus observability in software delivery
- 6:31Using the scientific method to improve developer experience
- 8:52Analyzing common bottlenecks and performance issues
- 12:31Impact of observability on DORA metrics and security
- 13:28Final summary and call to action
related talks
27mDeveloper Experience: It’s Time to Start Complaining
30mAndroid Developer Productivity
40mDeveloper Productivity Engineering for Open-Source Communities
11mAre You Ignoring Your Most Expensive Bazel Build Problems? (Develocity Can Help)
37mBecoming and Staying a Productive Developer with Build Scans, Build Validation Scripts and Gradle
36mDeveloper Productivity Engineering: What's in it for me?
55mAre your tests slowing you down?
51mAre Your Test Slowing You Down?