We are observing crashes in our customers environments with OldObjectSample events enabled.

The stacktrace looks like this:
---
resource_allocate_bytes(unsigned long, AllocFailStrategy::AllocFailEnum) (libjvm.so)
JfrStackTrace::JfrStackTrace() (libjvm.so)
JfrStackTraceRepository::record_for_leak_profiler(JavaThread*, int) (libjvm.so)
ObjectSampler::sample(HeapWordImpl**, unsigned long, JavaThread*) (libjvm.so)
JfrAllocationTracer::JfrAllocationTracer(Klass const*, HeapWordImpl**, unsigned long, bool, JavaThread*) (libjvm.so)
AllocTracer::send_allocation_in_new_tlab(Klass*, HeapWordImpl**, unsigned long, unsigned long, JavaThread*) (libjvm.so)
MemAllocator::allocate() const (libjvm.so)
InstanceKlass::allocate_instance(JavaThread*) (libjvm.so)
OptoRuntime::new_instance_C(Klass*, JavaThread*) (libjvm.so)
---

We have also the JVM flags that are used for one of those services:
---
 -XX:+UseCompressedOops -XX:+UseParallelGC -XX:GCTimeRatio=999 -XX:NewRatio=1 -XX:ParallelGCThreads=32 -Xlog:gc=info::utctime,uptime,level,tags -XX:+CrashOnOutOfMemoryError -XX:ReservedCodeCacheSize=256m -XX:+UnlockDiagnosticVMOptions -XX:+ShowHiddenFrames -XX:+DebugNonSafepoints -XX:-RewriteBytecodes -XX:MaxMetaspaceSize=256M -XX:-TieredCompilation -XX:ReservedCodeCacheSize=64M -XX:InitialCodeCacheSize=64M -XX:CICompilerCount=2 -XX:CompilationMode=high-only -XX:+UseAESCTRIntrinsics -XX:ActiveProcessorCount=32 -Xmx63488m -Xms63488m -XX:+UseTransparentHugePages -XX:NativeMemoryTracking=summary -XX:+PrintNMTStatistics
---