Handle possible ThreadMXBean findDeadlockedThreads UnsupportedOperationException

[Doug Hoard]

While testing the JMX Exporter against the latest JVM versions...

https://github.com/prometheus/jmx_exporter/blob/main/integration_test_suite/integration_tests/src/test/resources/docker-image-names.all.txt

It seems that azul/prime:8, azul/prime:11, and azul/prime:17 have changed behavior and no longer support findDeadlockedThreads, throwing an UnsupportedOperationException

java.lang.UnsupportedOperationException: Monitoring of Synchronizer Usage is not supported.
at java.management/sun.management.ThreadImpl.findDeadlockedThreads(ThreadImpl.java:431)
at io.prometheus.jmx.shaded.io.prometheus.client.hotspot.ThreadExports.addThreadMetrics(ThreadExports.java:98)
at io.prometheus.jmx.shaded.io.prometheus.client.hotspot.ThreadExports.collect(ThreadExports.java:175)
at io.prometheus.jmx.shaded.io.prometheus.client.hotspot.ThreadExports.collect(ThreadExports.java:169)

This exists in the pre-1.0.0 code, but will also cause issues in the 1.0.0 code.

The code below needs to be changed to handle such scenarios since the JVM is not required to support findDeadlockedThreads

client_java/prometheus-metrics-instrumentation-jvm/src/main/java/io/prometheus/metrics/instrumentation/jvm/JvmThreadsMetrics.java

Lines 102 to 107 in a8902cb

 if (!isNativeImage) { 
     GaugeWithCallback.builder(config) 
             .name(JVM_THREADS_DEADLOCKED) 
             .help("Cycles of JVM-threads that are in deadlock waiting to acquire object monitors or ownable synchronizers") 
             .callback(callback -> callback.call(nullSafeArrayLength(threadBean.findDeadlockedThreads()))) 
             .register(registry);