Uploaded image for project: 'OCP Technical Release Team'
  1. OCP Technical Release Team
  2. TRT-1559

Investigate why prow job loading is often very slow


    • Icon: Story Story
    • Resolution: Done
    • Icon: Normal Normal
    • 4.16.0
    • None
    • None
    • False
    • None
    • False

      In looking at a component readiness test page we see some failures that take a long time to load:

      https://prow.ci.openshift.org/view/gs/test-platform-results/logs/periodic-ci-openshift-release-master-nightly-4.15-e2e-metal-ipi-ovn-dualstack/1758641985364692992 (I noticed that this one resulted in messages asking me to restart chrome)

      We'd like to understand why it takes a long time to load these jobs and possible take some action to remediate as much of that slowness as possible.

      Taking a long time to load prow jobs will make our TRT tools seem unusable and might make it difficult for managers to inspect Component Readiness failures which would slow down getting them resolved.

      Some idea of what to look at:

      • see if the file size of the jobs is any bigger now than before esp. for runs with a lot of failures
      • see if the recent change that cuts the size of the intervals down is still working as expected
      • compare the file size of a passing run vs. one with a lot of failures

            dperique@redhat.com Dennis Periquet
            dperique@redhat.com Dennis Periquet
            0 Vote for this issue
            1 Start watching this issue