You have a high-traffic application running behind a load balancer with clients that are very sensitive to latency. How should you determine which back-end Amazon Elastic Compute Cloud application instances are causing increased latency so that they can be replaced?

A. By using the Elastic Load Balancing Latency CloudWatch metric.
B. By using the HTTP X-Forwarded-For header for requests from the load balancer.
C. By running a distributed load test to the load balancer.
D. By using the load balancer access logs.