SAP-C02 Question 469
Select 2An e-commerce company is running a critical order processing application on Amazon ECS using the Fargate launch type. Users have reported intermittent failures during order placement. Upon investigation, you find that the service scales down under low traffic, causing delays in processing new requests until the containers scale back up. How can you improve the reliability of the application while keeping operational overhead minimal?
- A
Configure an Application Load Balancer health check to determine unhealthy containers and automatically restart them.
- B
Enable ECS Service Auto Scaling with a higher minimum task count to ensure adequate resources are always available.
- C
Use AWS Lambda for request processing instead of Amazon ECS to avoid the need for container management.
- D
Implement a scheduled scaling policy in ECS to increase the task count during peak traffic hours.
- E
Enable ECS task draining functionality to ensure graceful shutdown of tasks during scaling events.
Show answer and explanation
Correct answers: B, D
Explanation
To improve reliability, you need to address the delays caused by reactive scaling. Increasing the minimum task count with ECS Service Auto Scaling ensures a baseline level of availability, while a scheduled scaling policy allows the service to proactively scale during anticipated traffic spikes. Together, these strategies minimize downtime and improve the reliability of the application.
- A. Incorrect.
Health checks will help detect unhealthy containers but do not address the issue of scaling down and subsequent delays in processing new requests.
- B. Correct.
ECS Service Auto Scaling with a higher minimum task count ensures that a baseline number of tasks are always running, thus improving reliability by reducing the scaling delay.
- C. Incorrect.
Switching to AWS Lambda might solve the scaling issue, but it introduces a complete architectural change, which is not necessary in this case.
- D. Correct.
A scheduled scaling policy ensures that the service scales proactively during expected traffic peaks, reducing reliance on reactive scaling during high traffic periods.
- E. Incorrect.
Task draining functionality ensures graceful termination of tasks but does not prevent the initial problem of scaling delays.