One full-stack engineer is developing an application to process images and videos from a highly scaled website. The incoming web traffic is very unbalanced and unstable during the day. When the web traffic rises up, it may require a maximum of 10 m4.xlarge EC2 instances. However, the requirement may drop down to 1 m4.xlarge instance. The processing for the application takes time, resulting in the delay of web application response. He prefers to deploy the App in Elastic Beanstalk. Which option should the engineer choose?

Answer options:

A.Deploy the application in one Elastic Beanstalk web server environment. Make sure that the environment uses a load-balancing, autoscaling configuration. Therefore, it can auto-scale depending on the traffic level.
B.Deploy the application in one Elastic Beanstalk worker environment. Create an SQS queue and an SQS daemon in EC2 to handling with the queue messages. Make sure that the environment uses a load-balancing and autoscaling configuration.
C.Create one Elastic Beanstalk worker environment and deploy the application in the new environment. Make sure that the application performs the short-running task in response to the POST from SQS daemon. Uses an auto-scaling configuration for the environment.
D.Create an Elastic Beanstalk web server environment and add the application bundle to it. Make sure that the environment uses an autoscaling configuration with a maximum at 10 and a minimum at 1.

Correct Answer – B
 
Option A is INCORRECT because the traffic is very unbalanced. Since it is a "long-running process" an Elastic Beanstalk Worker environment is preferred over the Elastic Beanstalk Web environment.
Option B is CORRECT because it is a "long-running process" an Elastic Beanstalk Worker environment is preferred here. Also, SQS has the following capabilities: Separating query from long-running task. Also, we can "park" the video and let CPUs process from the queue because we cannot wait on the video to be done before sending a response to the user and for building backlogs (Dead Letter Queues for later processing)
Option C is INCORRECT because it should be a long-time running task instead of a short-time running task.
Option D is INCORRECT because since the traffic is very unbalanced, we need to consider using an elastic load balancer for the same and this option does not use it.