Broker logs erroneously include the Artemis audit

[k-wall]

Describe the bug

Broker logs are now including the Artemis audit. The intent was that this was excluded, for performance reasons. This is a regression.

To Reproduce Steps to reproduce the behavior:

1. Create a queue address and then examine the broker's pod log.

Expected behavior No audit message

Screenshots If applicable, add screenshots to help explain your problem.

Additional context Add any other context about the problem here.

[k-wall]

I see that the Broker is loading the JVM's default logging.properties of the JVM rather than the logging.properties provided the broker-plugin. I have not yet worked out what caused the regression.

[k-wall]

The broker is given a Jboss logging properties file via a command line property -Dlogging.configuration=file:/var/run/artemis/split-1//serverData/etc//logging.properties

I confirmed that the suppression of audit messages was working at 0.29.2. I did this by oc rsh into the broker pod, editing /var/run/artemis/split-1//serverData/etc//logging.properties to reenable auditing. The logging.properties is reloaded dynamically (AMQ221056 :Reloading configuration: logging seen) and audit messages seen immediately.

The regression occurred at 0.30. The Broker is still consuming the logging.properties file, but for some reason it (or more likely the JBoss Logging Manager), is failing to act on all of the directives. It is odd because the logging format is still applied, but the rules for .audit.base and .audit.message are being ignored.

I noticed the following message on startup: WARNING: Failed to instantiate LoggerFinder provider; Using default.

This lead me to https://issues.redhat.com/projects/LOGMGR/issues/LOGMGR-254

I suspect that this issue is somehow related but I've been unable to identify the root cause.

[BobClaerhout]

I don't know whether my issue is related to this one but I'm commenting here anyway. Maybe my comments can contribute to finding the root cause of this. The issue I'm currently experiencing it the inability to configure the logging whatsoever. My goal is to have less logging of the broker since this is way too much for us. I'm currently using Enmasse version 0.31.
Things I've tried:

• Changing the JAVA_OPTS environment variable to configure the log level. This didn't work. Later on, I discovered following line in /opt/apache-artemis/custom/bin/env.sh: export JAVA_OPTS=" -Djava.security.egd=file:/dev/./urandom -Djavax.net.ssl.keyStore=/var/run/artemis/split-1//serverData/etc/enmasse-keystore.jks -Djavax.net.ssl.keyStorePassword=enmasse -Djavax.net.ssl.trustStore=/var/run/artemis/split-1//serverData/etc/enmasse-truststore.jks -Djavax.net.ssl.trustStorePassword=enmasse". To me, this looks like the JAVA_OPTS env variable is not used at all?
• After that, I also tried configuring the log level directly in the config file. Changing this resulted in an automatic reload of the configuration with the described log line by @k-wall. However, the log level was not changed.
• I tried rebooting the pod and was hoping the configuration was properly loaded at startup but this didn't work either. I discovered the logging configuration was overwritten with the default configuration. I was hoping for the AMQ_RESET_CONFIG env variable to disable the config override but that didn't work either although this was logged properly at the start.