Controllers not accessible after Upgrading past 2.277.4.2 when using CloudBees CI with platform EKS

Issue

I have deployed CloudBees CI in EKS with .OperationsCenter.platform="eks" and using the AWS Load Balancer Ingress Controller
After upgrading to 2.277.4.2 and stopping / starting the Controllers, I cannot access them anymore. They are unresponsive at $CONTROLLER_URL/operations-center-sso-navigate/?from=$CONTROLLER_PATH.

Environment

CloudBees CI (CloudBees Core) on Modern Cloud Platforms >= 2.277.4.2
AWS Load Balancer Controller < 2.2.0
Master Provisioning Plugin >= 2.5.28

Related Issue(s)

BEE-382: Update controller-provisioning-plugin to support Ingress in networking.k8s.io/v1
aws-load-balancer-controller #1772 add support for Ingress pathType

Explanation

Starting from version 2.277.4.2 of CloudBees CI, Controller ingresses are using the networking.k8s.io/v1 if supported by the underlying cluster. Instead of using 2 rules per controller /<controllerName>/ and /<controllerName>/*, it now uses a single rule on /<controllerName>/ with Prefix as pathType.

However, the Ingress API version networking.k8s.io/v1 is only supported by the AWS Load Balancer Controller since version 2.2.0. In earlier version, the single rule is not interpreted as a prefix but as an exact match and any access to /<controllerName>/* is not forwarded.

Resolution

Upgrade the AWS Load Balancer Controller to version 2.2.0 or later

This article is part of our Knowledge Base and is provided for guidance-based purposes only. The solutions or workarounds described here are not officially supported by CloudBees and may not be applicable in all environments. Use at your own discretion, and test changes in a safe environment before applying them to production systems.