If you are having trouble with Elastic Network Interfaces (ENIs) stuck in an “in-use” state after deleting your Bedrock AgentCore runtime, it can block you from cleaning up related resources like security groups, subnets, or CloudFormation stacks. Here’s a simple plan to resolve this issue:
First, give some time for the automatic cleanup process. AWS-managed ENIs usually get released shortly after the parent resource is deleted. Sometimes, this can take a bit longer—up to several hours—due to internal operations. If it’s been more than a few hours, proceed with the next steps.
Next, check that you’ve fully deleted the Bedrock AgentCore runtime. You can do this by logging into the Bedrock console or using the AWS Command Line Interface (CLI) to list your agent runtimes. Confirm that the resource no longer exists.
You should also verify if there are other related resources, like Elastic File System (EFS) mount target ENIs. Sometimes, these ENIs are linked to an EFS file system that’s still active. Make sure no EFS resources are still in use, as they could be holding onto those network interfaces.
If the ENIs are still stuck and you can’t delete them manually—because they are managed by AWS and not directly removable—you should reach out to AWS Support. You can do this by visiting the AWS Support Center. Provide them with details such as the ENI IDs, the ARN of the deleted runtime, and the date you performed the deletion. This information will help the support team locate the problem quickly and assist in releasing the ENIs.
Normally, AWS-managed ENIs should be cleaned up within 15 to 30 minutes after the related resource is deleted. If you see them lingering beyond that, support can help ensure they are released without delay.
Following these steps should help you clear up the lingering ENIs and get back to normal resource cleanup.