Ingress Nginx pod logs shows error when using .OperationsCenter.Platform=eks

Article ID:360058867672
1 minute readKnowledge base

Issue

  • After upgrading CloudBees CI to version 2.263.2.2, the Ingress Nginx pod logs shows several errors like the following:

nginx-ingress-nginx-controller-1q2w3e4r5t-q1w2e controller W0316 20:46:54.149733 7 controller.go:863] Error obtaining Endpoints for Service "<namespace>/ssl-redirect": no object matching key "jenkins-cloudbees/ssl-redirect" in local store
nginx-ingress-nginx-controller-1q2w3e4r5t-q1w2e controller W0316 20:46:54.149767 7 controller.go:863] Error obtaining Endpoints for Service "<namespace>/root-redirect": no object matching key "jenkins-cloudbees/root-redirect" in local store

Explanation

Since version 2.263.2.2, CloudBees CI in Modern Cloud Platform supports the AWS Application Load Balancer (ALB) Ingress controller and becomes the recommended solution when deploying CloudBees CI in EKS. Starting from this version, if the value eks is used for the helm attribute .Values.OperationsCenter.Platform, the CloudBees CI deployment automatically injects ALB Ingress rules and annotations. If the Kubernetes Ingress Nginx Controller is used instead of the ALB Ingress Controller, those rules cannot be read and produce those errors in the logs:

OperationsCenter: Platform: "eks"

Those errors are red herring and should not impact operations. It might impact the Ingress Controller on start up or controller restarts depending on the number of controllers.

Resolution

The recommended solution is the use the AWS Application Load Balancer Ingress Controller as documented at the EKS Install Guide.

Workaround

A workaround is to use the value aws for the helm attribute .Values.OperationsCenter.Platform:

OperationsCenter: Platform: "aws"