SAP-C02 Question 492
Select 2A company hosts a web application on AWS that experiences unpredictable traffic spikes. The application is currently deployed using an Auto Scaling group for the EC2 instances, fronted by an Application Load Balancer (ALB). The company wants to ensure minimal latency during traffic spikes while keeping costs optimized. Which of the following scaling strategies should you implement to meet these requirements?
- A
Configure a target tracking scaling policy that adjusts capacity based on average CPU utilization.
- B
Enable predictive scaling for the Auto Scaling group to scale out before anticipated traffic spikes.
- C
Use a scheduled scaling policy to add instances at fixed times during the day.
- D
Configure the ALB to use sticky sessions to reduce the load on back-end EC2 instances.
- E
Implement a step scaling policy to add a predefined number of instances when CPU utilization crosses a threshold.
Show answer and explanation
Correct answers: A, B
Explanation
To handle unpredictable traffic spikes effectively, both target tracking scaling and predictive scaling are ideal approaches. Target tracking scaling ensures that the environment dynamically adjusts to maintain performance based on metrics like CPU utilization, while predictive scaling anticipates traffic demands and scales out proactively. These methods together optimize performance and cost. Scheduled scaling and step scaling are less suited for unpredictable traffic, and sticky sessions do not address scaling needs.
- A. Correct.
This is a correct option. A target tracking scaling policy adjusts capacity dynamically based on metrics (e.g., CPU utilization), ensuring that the application can handle traffic spikes while maintaining performance and optimizing costs.
- B. Correct.
This is a correct option. Predictive scaling uses historical and forecasted data to anticipate traffic spikes and proactively scale out, which helps minimize latency during traffic surges.
- C. Incorrect.
This is incorrect. Scheduled scaling is based on fixed time intervals and is not effective for unpredictable traffic spikes, as it cannot account for variability in traffic patterns.
- D. Incorrect.
This is incorrect. While sticky sessions can help reduce the load on the back-end by reusing the same instance for a user session, it does not address the need to scale out dynamically during traffic spikes.
- E. Incorrect.
This is incorrect. Step scaling adds a fixed number of instances when thresholds are breached, but it may not be as responsive or cost-efficient as target tracking or predictive scaling in handling unpredictable traffic patterns.