• Type: Improvement
    • Status: Resolved
    • Priority: Blocker
    • Resolution: Fixed
    • Affects Version/s: 2.0.1
    • Fix Version/s: 4.0.0
    • Component/s: app.jobbrowser
    • Labels:
    • Target Version:
    • Epic Link:
    • Sprint:
      March 7-11, 2016, March 14-18, 2017


      Goals, Unify and aggregate:


      • Jobs :YARN, Impala, Spark, Sqoop....
      • ATS Flows
      • Workflows
      • Schedules
      • Bundles

      App stats:


      • Batch jobs: query export, S3 copies, indexing...
      • Scheduled jobs: integrated schedules

      Some ideas about general UX:

      1. Job browser displays only the first 1000 jobs from resource manager start
      2. pagination with 1000s of jobs
      3. log pagination as we break
      4. kill wf of hs2 script, error when stopping hs2 in jb
      5. pagination, job status filtering is broken after each background refresh
      6. Cannot kill ‘Accepted’ jobs from jobbrowser as can't ckick on them
      7. re-think how we can get the job logs in a stable way + testable
      8. link back to objects, like Hive, Sqoop... queries
      9. better spark
      10. charting summary of # of tasks, progress bars...
      11. breadcrumbs
      12. speed with 10k+ jobs
      13. aggreate hive, impala, oozie jobs?
      14. Mappers and Reducers show the same progress
      15. tabs don't always support back button
      16. progress label color
      17. left bar should have hierarchical links (jobs --> task --> ...)
      18. left bar should have more options/icons
      19. tasks: put the task in green/orange/red depending on its status maybe
      20. Job priority is gone on the application lists :
      21. Task bottleneck visualization
      22. Timeline view?
      23. Smarter filtering
      26., queues, tags...
      27. Spark
      28. We could live progress the tasks, counters... when on their respective tabs
      29. Not super clear facet selection (succeeded checked)

      Large issues
      Random errors when killing running jobs, or jobs that just finished
      Tried killing the hive job which was 50% complete - it gave error message “There was a problem communicating with the server. Refresh the page”

      Logs for running jobs fails with exception
      Sometimes only


          Issue Links



              • Assignee:
                romain Romain Rigaux
                romain Romain Rigaux
              • Votes:
                0 Vote for this issue
                2 Start watching this issue


                • Created: