The benchmark that changed my opinion about Jakarta EE in 2026

A revised benchmark compared Spring Boot, Payara Micro, and Embedded GlassFish on the same PostgreSQL-backed Jakarta EE API.

After standardizing the JDK, heap, pools, warmup, and measurement windows, the results became more controlled and less one-sided.

Spring Boot emerged as the best default choice for new projects, while Payara Micro and Embedded GlassFish remain viable in existing Jakarta EE stacks.

The article highlights how benchmark methodology can materially change conclusions, especially around latency, throughput, RSS, and GC behavior.