[EKS] [bug]: Managed Node Groups with Warm Pools → Add-ons in "Degraded" Status

[maximeroma]

Context

AWS recently introduced support for warm pools for EKS Managed Node Groups.

Since enabling this feature, we’ve observed unexpected behavior affecting EKS add-ons.

• [EKS] [request]: Managed Node Group support for ASG warm pools #1387
• https://aws.amazon.com/fr/about-aws/whats-new/2026/04/amazon-eks-managed-node-groups-ec2-warm-pools/

Observed Behavior

After enabling warm pools on a managed node group:

Most EKS add-ons switch to DEGRADED status
Exception: CoreDNS remains in ACTIVE status
Workloads appear to run normally, but add-on health status is incorrect

Expected Behavior

Remain in ACTIVE status when functioning correctly
Accurately reflect the real health state of the cluster

Actual Behavior

Affected add-ons:
kube-proxy
VPC CNI
EBS CSI
Node monitor agent
Reported status: DEGRADED

Environment

EKS version: 1.35
Region: eu-west-3
Node OS: Bottlerocket

[knottnt]

Thank you for reporting this. Team is investigating.