How Netflix Uses Java - 2026 Edition
summary
Netflix runs about 3,000–4,000 Java services standardized on Spring Boot with DGS/federated GraphQL for frontend discovery and gRPC for high‑performance RPC; upgrades use a Gradle Jakarta bytecode transform, OpenRewrite and a Cloud Code–driven batch workflow to migrate Spring Boot versions; runtime choices shift to generational ZGC as default (cutting GC pauses and IPC errors), experimental virtual threads with context propagation workarounds, Leyden AOT experiments, and agentic workflows via Spring AI to analyze and act on profiling data.
AI-written summary · about
chapters (14)
- 0:00Start
- 2:37The Netflix Architecture
- 8:03Why Java?!
- 11:31Spring Boot
- 15:42Testing with Spring Boot
- 17:40Spring Boot version ipgrades
- 23:38Why upgrade at all?
- 26:23DGS - GraphQL
- 29:33The JDK
- 30:05ZGC - The New Default
- 34:39Virtual Threads
- 41:27Building Agentic Workflows
- 48:05Java vs. Python for GenAI
- 49:49Q&A
related talks
48mHow Netflix Uses Java
50mHow Netflix Is Upgrading to Spring Boot 3
28mHow Netflix Makes gRPC Easy to Serve, Consume, and Operate - Benjamin Fedorka, Netflix
01h 29mSpring Boot and Netflix OSS
30mThe Netflix Domain Graph Service Framework - OSS GraphQL for Spring Boot
59mJava microservices from Netflix OSS to Kubernetes
40mCloud Native Java
01h 27mSpring Cloud at Netflix