An eCommerce website is hosted on two EC2 instances that sit behind an Elastic Load Balancer. The response time of the website has slowed dramatically. The analytics team found that one of the EC2 instances failed, leaving only one instance running. What is the best course of action to prevent this from happening in the future?

Answer options:

A.Change the instance size to the maximum available to compensate for the failure.
B.Use CloudWatch to monitor the instances via VPC Flow Logs.
C.Configure the ELB to perform health checks on the EC2 instances and implement auto-scaling.
D.Replicate the existing configuration in several regions for failover.

Correct Answer: C
Option C is correct. Using the elastic load balancer to perform health checks will determine whether or not to remove a non- or underperforming instance and have the auto-scaling group launch a new instance.
Option A is incorrect because increasing the instance size doesn’t prevent the failure of one or both instances. Therefore the website can still become slow or unavailable.
Option B is incorrect because monitoring the VPC flow logs for the VPC will capture VPC traffic, not traffic for the EC2 instance. 
Option D is incorrect because replicating the same two instance deployment may not prevent the failure of instances and could still result in the website becoming slow or unavailable.
References:
https://media.amazonwebservices.com/AWS_Building_Fault_Tolerant_Applications.pdf
https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/flow-logs.html#working-with-flow-logs