AWS LBC recreated the ALB

[brnpimentel]

Describe the bug
The AWS Load Balancer Controller unexpectedly recreated an ALB and its associated Target Groups without any manual deployment or configuration changes.

Steps to reproduce

1. Deploy the AWS Load Balancer Controller in an EKS cluster.
2. Apply a working Ingress resource to create an ALB.
3. Observe the ALB creation and normal functionality.
4. Wait or monitor without making any changes.

Expected outcome
The ALB should remain stable and unchanged unless there are intentional updates to the Ingress or related resources.

Environment

• AWS Load Balancer controller version: v2.9.2
• Kubernetes version: 1.29.10
• Using EKS (yes/no), if so version? yes.

Additional Context:
Logs show that the controller deregistered targets, deleted listener rules, and recreated the ALB. There were no manual changes applied to the Ingress resource or ALB configuration. This issue impacts stability in production environments.

Has anyone experienced this?

[zac-nixon]

Did you remove the ingress class specification? The controller won't delete resources unless prompted.

[rick-reihl]

I am having the same issue. This is happening in multiple clusters and accounts. I tracked down one deployment that triggered the recreation and the only manifest difference is the docker image tag used. I tracked the cloudtrail logs and the only difference in the new alb is the security group, which has the same rules as the previous.

The ingress objects have been unchanged for 30+ days. There is no reason from my side this should have been recreated. Some clusters are unaffected with the same deployments happening - dev the alb is recreated, same deployment later test is fine, alb has never been recreated.

AWS Load balancer controller version: v2.7.2
Using EKS: yes - v1.29.10