Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-38087

initial search call extremely slow

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Duplicate
    • Icon: Major Major
    • blueocean-plugin
    • None

      After a restart of Jenkins, a refresh of the blueocean UI causes a call to the following endpoint:

      <Jenkins URL>/blue/rest/search/?q=type:pipeline;excludedFromFlattening:jenkins.branch.MultiBranchProject,hudson.matrix.MatrixProject&filter=no-folders&start=0&limit=26

      The very first time this runs it is SUPER slow - irrespective of how long the instance has been up and running.

      I have a 60s timeout on my http frontend - and it hits this and the frontend returns a 504.

      The second time I reload the UI, it takes ~12s and return 11kB of information (the instance is relatively small).

          [JENKINS-38087] initial search call extremely slow

          James Dumay added a comment -

          Sounds like a cold cache problem that is innate to Jenkins. Moving this out of the sprint to the bottom of the backlog.

          James Dumay added a comment - Sounds like a cold cache problem that is innate to Jenkins. Moving this out of the sprint to the bottom of the backlog.

          Michael Neale added a comment -

          jamesdumay Part of it is warming up, but 12s is still too slow when warmed up (due to reasons mentioned on originating ticket) so I think I will open another ticket to track the persistent slowness and lack of back end pagination

          Michael Neale added a comment - jamesdumay Part of it is warming up, but 12s is still too slow when warmed up (due to reasons mentioned on originating ticket) so I think I will open another ticket to track the persistent slowness and lack of back end pagination

          Michael Neale added a comment -

          bwalding vivek has a fix in progress on the linked ticket - once that is in (he is verifying with dogfood workspace locally) - might get you to try again, and re-open this if it is still a problem.

          Basically pagination was still causing full calculations, thrashing disk. This would make things worse after warm up naturally so hopefully it will address both the unacceptably slow load for page of data, and the warm up.

          Michael Neale added a comment - bwalding vivek has a fix in progress on the linked ticket - once that is in (he is verifying with dogfood workspace locally) - might get you to try again, and re-open this if it is still a problem. Basically pagination was still causing full calculations, thrashing disk. This would make things worse after warm up naturally so hopefully it will address both the unacceptably slow load for page of data, and the warm up.

            Unassigned Unassigned
            bwalding Ben Walding
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated:
              Resolved: