Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-76496

[Docs bug] Performance issue with Ingress on v4.17

XMLWordPrintable

    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • None
    • Critical
    • None
    • None
    • Rejected
    • None
    • Customer Escalated
    • In Progress
    • Bug Fix
    • Hide
      This bugfix adds ClosedClientConnectionPolicy field to the IngressController API, enabling configuration of option abortonclose setting on the router. By exposing this option, users can fine-tune how the router handles TLS operations. This helps mitigate excessive CPU usage observed when large RSA keys (>2048 bits) are used under heavy load.
      Show
      This bugfix adds ClosedClientConnectionPolicy field to the IngressController API, enabling configuration of option abortonclose setting on the router. By exposing this option, users can fine-tune how the router handles TLS operations. This helps mitigate excessive CPU usage observed when large RSA keys (>2048 bits) are used under heavy load.
    • None
    • None
    • None
    • None

      Description of problem:

          The customer reported poor performance and connectivity issues with their application after upgrading to OpenShift Container Platform (OCP) version 4.17. The issue is characterized by a 30% success rate with many missing packets during load testing using the `vegeta` benchmark tool, compared to a 99.9% success rate in the 4.12 environment. The application is exposed using MetalLB with a LoadBalancer type service. The business impact is the inability to upgrade 4.12 clusters to 4.17 due to this performance degradation.

Configuration:

The issue is consistently reproduced every time the test is run on the 4.17 router.
The customer uses an external load balancer, MetalLB, with the "Cluster" load-balancing method.
Significantly higher CPU usage
Worse performance due to the overhead of TLS termination

      Actual results:
      Having an issue some the requests are getting failed partially

      Expected results:
      Same test working fine without any issue on 4.12.35

              alebedev@redhat.com Andrey Lebedev
              rhn-support-klakhwar Ketan Lakhwara
              Ketan Lakhwara
              None
              Hongan Li Hongan Li
              None
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Created:
                Updated:
                Resolved: