SAP-C02 Question 278
Select 4Your company uses an Auto Scaling group for a fleet of EC2 instances that handle a web application with highly variable traffic. To optimize cost and performance, you need to ensure that the Auto Scaling group responds effectively to both sudden spikes in traffic and gradual changes in demand. Which combination of scaling policies would achieve this goal?
- A
Use a target tracking scaling policy to maintain a specific average CPU utilization.
- B
Use a scheduled scaling policy to scale out during peak traffic hours and scale in during off-peak hours.
- C
Use a step scaling policy to respond to changes in a CloudWatch alarm threshold with specific scaling actions.
- D
Use a predictive scaling policy to forecast demand and provision capacity ahead of time.
- E
Use a manual scaling policy to manually adjust the desired capacity based on observed traffic patterns.
Show answer and explanation
Correct answers: A, B, C, D
Explanation
To optimize cost and performance for an application with highly variable traffic, a combination of scaling policies is typically required. Target tracking handles gradual changes, scheduled scaling addresses predictable demand patterns, step scaling reacts to sudden traffic spikes, and predictive scaling provides proactive capacity adjustments based on forecasts. Manual scaling, while useful in specific cases, is not suitable for dynamic environments requiring automated responses.
- A. Correct.
Correct: Target tracking scaling policies are ideal for maintaining a specific utilization metric, such as CPU utilization, and can handle gradual changes in demand effectively.
- B. Correct.
Correct: Scheduled scaling is a cost-effective way to handle predictable traffic patterns, such as peak hours, by pre-scheduling scaling actions.
- C. Correct.
Correct: Step scaling policies allow for more precise control when reacting to CloudWatch alarms, making them suitable for handling sudden traffic spikes.
- D. Correct.
Correct: Predictive scaling uses machine learning to forecast demand, enabling the Auto Scaling group to proactively adjust capacity and avoid latency during traffic surges.
- E. Incorrect.
Incorrect: Manual scaling is not practical for highly variable traffic as it requires constant human intervention and does not provide the automated response needed in this scenario.