A company must capture and log traffic for Nitro-based EC2 instances. The network team enabled VPC Traffic Mirroring to send mirrored traffic to a second Auto Scaling group of EC2 instances and put a Network Load Balancer (NLB) in front of those targets, but no mirrored traffic reaches the EC2 instances behind the NLB. How should the team configure traffic mirroring to use the NLB endpoint?

VPC Traffic Mirroring sends mirrored traffic as encapsulated GRE packets. Therefore, the target for this traffic must be able to receive GRE packets. A Network Load Balancer (NLB) can be used as a target for Traffic Mirroring, but it must be configured with a UDP listener to properly receive and forward the GRE-encapsulated traffic to its registered targets. Selecting the NLB as a source is incorrect because Traffic Mirroring sources are network interfaces (ENIs), not load balancers. Using a TCP listener is incorrect because GRE is a Layer 3 protocol, and its encapsulation is typically handled over UDP for load balancing purposes, not TCP. While GRE itself doesn't use UDP or TCP, NLBs require a listener protocol, and UDP is the appropriate choice for forwarding GRE-encapsulated traffic in this scenario.