javamelody | JavaMelody : monitoring of JavaEE applications | Monitoring library

As of Spring Boot 2.4, JUnit 5’s vintage engine has been removed from spring-boot-starter-test. If we still want to write tests using JUnit 4, we need to add the following Maven dependency:

JPA monitoring is completely different than SQL monitoring: do not try net.bull.javamelody.jpa.provider to monitor SQL. And it is useless to try hibernate.connection.driver_class if you use a datasource from JNDI.

For SQL monitoring, you can try to monitor SQL from the datasource in JNDI:

• either add a javamelody parameter datasource, for example with a system property: -Djavamelody.datasources=java:jboss/datasources/myDS
• or rename your datasource to java:jdbc/myDS

Then check the "Debugging logs" at the bottom of the javamelody reports.

And because you are using an EAR, see https://github.com/javamelody/javamelody/wiki/UserGuideAdvanced#setup-javamelody-in-an-ear-file

Looking at the stacktrace the failure is happening for com/uwaytech/modules/tdr/ctrl/TdrCtrl$1 which seems to be an anonymous class inside TdrCtrl. can you check the below link if it helps.

https://stackoverflow.com/a/11035641/11207493

There was an escape issue in this part of the docs. It now says:

The parameter storage-directory is the name of the directory of storage (javamelody by default). If the name of the directory starts with '/' (or on Windows, with drive specifier followed by '\', or if its prefix is "\\"), it is considered as an absolute path, otherwise it is considered as relative to the temporary directory

So you can write in your context:

Near the start of the Maven output, I see this line:

I could not find the solution to monitor using url and user name and password.

I converted application to use jndi connection pool and able to monitor.

Per the log file, there are more than 10,000 started threads. That's a lot even if we don't look at the less that 2 CPUs/cores reserved for the container (limits.cpu = request.cpu = 1600 millicores).

Each thread, and its stack, is allocated in memory separate from the heap. It is quite possible that the large number of started threads is the cause for the OOM problem.

The JVM is started with the Native Memory Tracking related options (-XX:NativeMemoryTracking=detail, -XX:+UnlockDiagnosticVMOptions, -XX:+PrintNMTStatistics) that could help to see the memory usage, including what's consumed by those threads. This doc could be a starting point for Java 11.

In any case, it would be highly recommended to not have that many threads started. E.g. use a pool, start and stop them when not needed anymore...

This was so annoying to debug due to not knowing the origin of the error. None of the jenkins logs were helpful between FINEST and SEVERE. After working with our lead of devops for a couple days we couldn't find anything on the system or server settings that changed the behavior.

I ran across someone who had a similar issue on GitHub saying that for one of their plugins, it would break if there was an extra EOL symbol. That got me thinking, could this be a parsing error for plugin input?

First I disabled everything that wasn't a build step. This fixed nothing. Then I started removing one plugin type at a time in the Build Step section and eventually narrowed it down to Windows Batch Command build steps. Through trial and error I determined it was 1 of 3 batches and debugged it down to finally a single line. This was the part of the batch where it was hidden:

It looks like Tomcat 8.5.50 contains a bunch of regression fixes in Catalina with resource loading which may be affecting the loading of Clojure source files. Suggest either downgrading to same as the working one or upgrading to 8.5.50.

https://tomcat.apache.org/tomcat-8.5-doc/changelog.html