EC2 instances should not use multiple ENIs
Flags EC2 instances using multiple Elastic Network Interfaces, which can create complex network configurations and unintended routing.
- service
- EC2
- severity
- Low
- scope
- evaluated in every region
How to fix it
Detach the extra network interfaces so the instance has one. Only secondary interfaces can be detached, the primary one stays for the life of the instance, and anything bound to the addresses you remove stops working at once. Two populations fail by design and are better suppressed than fixed: nodes in an EKS cluster, where the VPC CNI attaches an interface per batch of pod addresses, and HPC instances using Elastic Fabric Adapters.
In the console
EC2 → Instances → the instance → Actions → Networking → Manage network interfaces
With the AWS CLI
aws ec2 detach-network-interface --attachment-id <eni-attachment-id> --region <region>
See which of your accounts this check is failing on.
Book a demo