Uploaded image for project: 'Flink'
  1. Flink
  2. FLINK-16721

getProcessCpuLoad seems to report a wrong cpu load

    XMLWordPrintableJSON

    Details

    • Type: Improvement
    • Status: Closed
    • Priority: Major
    • Resolution: Cannot Reproduce
    • Affects Version/s: 1.10.0
    • Fix Version/s: None
    • Component/s: Runtime / Metrics
    • Labels:
      None

      Description

      I am monitoring the flink_taskmanager_Status_JVM_CPU_Load in the prometheus. As there is some random spikes (see attachment in cpu trend), but when i use some tool to get the thread load(using the analytical tool: https://github.com/oldratlee/useful-scripts/blob/master/show-busy-java-threads) , I didn't see any abnormal(see busythreads attachment) 

       

      And then I found the issue about mxBean::getProcessCpuLoad in stackoverflow. https://stackoverflow.com/questions/38421242/operatingsystemmxbean-getsystemcpuload-and-getprocesscpuload-values

       some one suggest the metric can be inaccurate in some cases. 

      Can anybody help me understand what's going on in my case?

        Attachments

        1. busythreads.txt
          216 kB
          xiaogang zhou
        2. cpu trend.png
          50 kB
          xiaogang zhou

          Activity

            People

            • Assignee:
              Unassigned
              Reporter:
              zhoujira86 xiaogang zhou
            • Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: