RFR: 8325616: JFR ZGC Allocation Stall events should record stack traces

Stefan Karlsson stefank at openjdk.org
Mon Feb 12 10:47:12 UTC 2024


On Mon, 12 Feb 2024 10:40:08 GMT, Stefan Karlsson <stefank at openjdk.org> wrote:

> We currently don't record the stack traces when sending the JFR ZGC Allocation Stall events. Let's add stack traces to the the jdk.ZAllocationStall event.

I'm running targeted stress testing on this and will run this through tier1-tier3 + tier4-tier7 on Linux.

Output from `bin/jfr print`:

jdk.ZAllocationStall {
  startTime = 11:21:54.584 (2024-02-12)
  duration = 1.60 s
  type = "Small"
  size = 2.0 MB
  eventThread = "pool-3-thread-5" (javaThreadId = 360764)
  stackTrace = [
    java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.newConditionNode() line: 1622
    java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await() line: 1692
    java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take() line: 1177
    java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take() line: 899
    java.util.concurrent.ThreadPoolExecutor.getTask() line: 1070
    ...
  ]
}

-------------

PR Comment: https://git.openjdk.org/jdk/pull/17803#issuecomment-1938426619
PR Comment: https://git.openjdk.org/jdk/pull/17803#issuecomment-1938431137


More information about the hotspot-jfr-dev mailing list