What’s New in Spring Boot 3.2? Explore Java 21 Features and Virtual Threads

Spring Boot 3.2 was officially released today, shortly after Java 21, and includes extensive optimizations for the new JDK.

New Features

The official blog outlines the following highlights:

Support for virtual threads (Project Loom)

Initial support for JVM Checkpoint/Restore (CRaC project)

SSL bundle hot‑reloading

Significant observability improvements

Support for RestClient

Support for JdbcClient

Support for Jetty 12

Apache Pulsar integration

SSL bundle support for Kafka and RabbitMQ

Redesigned nested JAR handling

Docker image build enhancements

For a full list of dependency upgrades, see the Spring Boot 3.2 release notes.

Video Overview

Josh Long’s 80‑minute video provides a hands‑on demonstration of the new capabilities.

[00:16] Spring Boot 3.2 introduces many new features, including virtual threads and Project Loom.

Java 21 is a key component of Spring Boot 3.2.

Java 21 adds sealed types, pattern matching, enhanced switch expressions, and records.

Java 21 is marketed as data‑oriented programming to improve performance of large monoliths.

[10:13] String formatting and multiline variables are now convenient, with new features such as records, sealed types, enhanced switch, and pattern matching (destructuring operators are still in development).

[20:27] Demonstration of a collection that skips duplicates, logs the current thread name, and sleeps for 100 ms, showcasing virtual thread usage.

[30:41] Building a SQL‑connected application using Java 21, Maven, and supporting libraries, including Testcontainers to start Docker images and Spring Boot starter JDBC for PostgreSQL.

[40:57] Declarative interfaces simplify code, providing a cat‑facts endpoint; new JDBC and REST clients can be used, and Project Loom integration offers consistent virtual thread support across layers.

[51:09] The SmartLifeCycle interface in Spring represents typical lifecycle methods, allowing custom start/stop management.

[01:01:23] Generation of unique key‑value pairs via a file and configuration of a self‑signed SSL certificate on port 8443.

[01:11:38] Using Docker Compose to launch a Zipkin instance, adding sampling probability in code, and leveraging AOP annotations to display trace and span IDs for distributed tracing.