Uploaded image for project: 'SonarQube'
  1. SonarQube
  2. SONAR-11455

Fix performance of issue persistence when processing an analysis report

    Details

    • Edition:
      Community
    • Production Notes:
      None

      Description

      This ticket is being created from two sources:

      perf issue reported by support related to a customer

      The pattern observed by support is the following:

      1. persist issues step is the costliest step
      2. Primo analysis creates and stores 1.9 millions issues, persist issues step takes 1640214ms.
      3. Second analysis inserts 7942 issues and updates 31138 issues, persist issues step takes 1182291ms, ie. 72% of the time of the step during the primo analysis.

      Similar situations have been observed by support on peach on project is LOC and issue size similar to the customer's project.

      perf issue reported dev looking at code

      Persist issue step uses a batch session to reduce the number of network round trips to the DB.
      While the batch session is efficiently leveraged during primo analysis as only inserts are performed, this is not the case at all during next analysis where inserts and updates are mixed.

      On top of it, a bug has been observed, tracked in SONAR-11454.

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                sebastien.lesaint Sebastien Lesaint
                Reporter:
                sebastien.lesaint Sebastien Lesaint
              • Votes:
                0 Vote for this issue
                Watchers:
                2 Start watching this issue

                Dates

                • Due:
                  Created:
                  Updated:
                  Resolved: