3 years of Quarkus in production, what have we learned?
summary
Quarkus migration of a decade-old monolith runs across ~9–10 scrum teams and delivers concrete practices: use a Maven build cache to drop small local builds from ~13s to 2.6s and large CI builds from ~20min to ~2–3min; prefer the Quarkus BOM for dependency versions; expect DevMode restarts of 3–4s for big modules and limit test profiles to avoid repeated restarts; native images cut memory to ~110MB vs JVM growth that dies under ~800MB; add a semantic-versioning plugin and slow-span filtering for usable tracing.
AI-written summary · about
chapters (11)
- 0:00Introduction and background
- 2:39Developer experience with Quarkus
- 7:37Moving to multimodule projects
- 11:55Build caching benefits
- 15:51Security in production deployments
- 18:23Memory management in JVM vs native
- 24:41Innersource and versioning challenges
- 30:59Quarkus updates and migration
- 35:15Testing pitfalls and solutions
- 39:51Optimizing tracing for performance
- 45:55Q&A session
related talks
44m3 Years of Quarkus in Production, What Have We Learned?
46mMigrating a real‑world application to Quarkus
51mQuarkus why, how and what
45mFaster, greener, and happier- why Quarkus should be your next tech stack
59mQuarkus for Spring Developers
46mQuarkus why, how and what
54mQuarkus: Java development turned into delight
34mLets be Real Quarkus in the Wild