Question

A company runs a production workload across three Availability Zones in a single VPC. Private subnets in all three AZs currently route outbound internet traffic through a single NAT gateway located in AZ-a, which is used for downloading OS patches and calling external APIs. During a recent AZ-a service disruption, workloads in AZ-b and AZ-c lost all outbound internet connectivity even though their own compute resources remained healthy. The DevOps team must eliminate this cross-AZ dependency with the least ongoing operational effort. What should they do?

Accepted Answer

Deploy one NAT gateway per AZ and configure each private subnet's route table to send 0.0.0.0/0 traffic to the NAT gateway in its own AZ.

Answer

Replace the NAT gateway with a fleet of self-managed NAT instances in an Auto Scaling group spanning all three AZs behind a Network Load Balancer.

Answer

Create a second NAT gateway in AZ-b and configure Route 53 health checks to fail traffic over from the AZ-a NAT gateway when it becomes unavailable.

Answer

Attach an egress-only internet gateway to the VPC and point all private subnet route tables to it for high availability across AZs.