Skip to main content
Interface VPC endpoints and Gateway Load Balancer endpoints are billed per hour in each Availability Zone they are deployed to, whether or not any traffic goes through them. nOps finds endpoints that have been billed for months without processing any data, and shows their hourly charge as the saving.

See it in the nOps Optimization Hub

Examples, savings and how nOps automates this recommendation.

Why it matters

Endpoints are often created for every AWS service in every VPC by a template, and many are never used. Each one is billed in every Availability Zone it covers.

How nOps finds it

nOps reads the last 90 days of billing for each VPC endpoint. One is flagged when its hourly charge was billed on at least 85 of those days and in the last 2 days, and it processed zero bytes in the whole 90 days. The saving is the hourly charge billed in the last 30 days, and must be at least $1.

Before you start

Permission to delete VPC endpoints (ec2:DeleteVpcEndpoints). Find out who owns the endpoint.

How to implement

  1. Review the endpoint’s service, type and private DNS setting: aws ec2 describe-vpc-endpoints --vpc-endpoint-ids ID.
  2. If private DNS is on, check that applications in the VPC can still reach the service through its public endpoint (or a NAT gateway) after deletion.
  3. Delete it: VPC → Endpoints → Actions → Delete VPC endpoints, or aws ec2 delete-vpc-endpoints --vpc-endpoint-ids ID.
  4. Remove it from Terraform or CloudFormation so it is not recreated.

Risks and considerations

  • With private DNS on, the service hostname resolves to the public endpoint after deletion. Instances without internet access lose access to that service.
  • Deletion cannot be undone, but you can create a new endpoint for the same service.