Uploaded image for project: 'AMQ Broker'
  1. AMQ Broker
  2. ENTMQBR-1918

[AMQ7, critical-analyzer] add extra detail to output for analysing cause of timeout


    • Icon: Story Story
    • Resolution: Done
    • Icon: Major Major
    • AMQ 7.2.2.GA
    • None
    • None
    • None
    • Release Notes
    • The logs for Critical Analyzer now provide extra information about a slow-performing broker when you enable tracing on the package named org.apache.activemq.artemis.utils.critical.
    • Documented as Feature Request
    • ARTEMIS-2157, ARTEMIS-2936
    • AMQ Broker 1842

      Currently the critical-analyzer does not print out detail with regards to which critical path has timed out. It does generate a thread dump but if the path has subsequently completed the critical path that timed out MAY not be listed in the thread dump.

      2018-08-06 08:52:46,834 ERROR [org.apache.activemq.artemis.core.server] AMQ224079: The process for the virtual machine will be killed, as component org.apache.activemq.artemis.core.io.buffer.TimedBuffer@421e361 is not responsive
      2018-08-06 08:52:47,309 WARN  [org.apache.activemq.artemis.core.server] AMQ222199: Thread dump: *******************************************************************************

      To debug we really need

      • a WARN message when a critical path has timed out.
      • 2 thread dumps rather than 1 therad dump with a small delay ( maybe 10 seconds) between each. It is difficult to surmise what threads are waiting/blocked for a significant period when only one thread dump is generated.
      • generate a full heap dump to analysis the contents of memory when the issue occurred. Perhaps that could be configurable option in the analyser in case some customers feel it is too heavy.

            csuconic@redhat.com Clebert Suconic
            rhn-support-pfox Patrick Fox (Inactive)
            0 Vote for this issue
            7 Start watching this issue