Uploaded image for project: 'Flink'
  1. Flink
  2. FLINK-26074

Improve FlameGraphs scalability for high parallelism jobs

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Closed
    • Major
    • Resolution: Fixed
    • None
    • 1.16.0
    • None

    Description

      The FlameGraph feature added in FLINK-13550 issues 1 RPC call per subtask. This may cause performance problems for jobs with high paralleism and a lot of subtask running on the same TaskManager. It should be possible to improve this by grouping thread sampling requests using 

      ThreadMXBean.getThreadInfo(long[] ids, int maxDepth)

      instead of the currently used individual calls

      ThreadMXBean.getThreadInfo(long id, int maxDepth)

      Attachments

        Issue Links

          causes

          Bug - A problem which impairs or prevents the functions of the product. FLINK-30239 The flame graph doesn't work due to groupExecutionsByLocation has bug

          • Major - Major loss of function.
          • Closed
          links to

          Web Link GitHub Pull Request #19228

          Activity

            People

              afedulov Alexander Fedulov
              afedulov Alexander Fedulov
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: