Before reporting an issue

[X] I have read and understood the above terms for submitting issues, and I understand that my issue may be closed without action if I do not follow them.

Area

dist/quarkus

Describe the bug

The class org.jgroups.protocols.TP memorizes the log level TRACE as it is started early before the logging system is initialized.

This leads to increased CPU usage and CPU. This is AFAIK a regression from #24860 / cc: @mabartos.

Version

main

Regression

[X] The issue is a regression

Expected behavior

No unnecessary messages should be formatted.

Actual behavior

The messages are formatted using the String formatter, which uses additional CPU cycles and memory.

This was not a problem in KC22, when the caches were initialized later.

How to Reproduce?

Start Keycloak's nightly in production mode, then set a breakpoint in the TP class at the trace messages.

Anything else?

I'll prepare a PR. Memorizing the trace level was an optimization in JGroups in https://issues.redhat.com/browse/JGRP-2130, I'll comment there to raise awareness as it is difficult to handle.