Hi, I have deployed an ECS Fargate using an existing VPC that has an on-premise gateway. The infrastructure was deployed using Terraform and, initially, the ECS was able to pull images from ECR. However, when I deployed the service using a GitHub action, the task was unable to pull images from ECR and timed out with an error message saying "unable to pull images from repository, API timeout." To fix this issue, I created a public NAT gateway associated with the VPC and created a NAT route to the private subnet.
However, I have noticed that other services deployed in AWS are now reporting huge network latency when they connect to the on-premise. Could this be happening as a result of adding the NAT?