SAA-C03 Practice Question 77: ECS Fargate Launch Type
AWS Certified Solutions Architect – Associate (SAA-C03) · Test 8 · question 7 of 10
A team has containerised a microservice and wants to run it on Amazon ECS. Traffic is bursty and hard to forecast, the team has no Linux administrators, and nobody wants to decide on instance types, patch container hosts, or think about how tasks pack onto a cluster. They are willing to pay a premium per unit of compute to avoid all of that.
Which ECS configuration fits?
Show answer and explanation
Answer:
Fargate lets you run containers on ECS without managing servers or clusters of EC2 instances. AWS states that it removes the need to choose server types, decide when to scale clusters, or optimise cluster packing, the three things the team explicitly does not want to do. You package the application, declare the CPU and memory each task needs, and launch it; each task runs on a patched platform version and gets its own isolation boundary. Paying a premium per unit of compute in exchange for that is precisely the trade-off stated in the question.
Option B is the launch type the team is trying to avoid. An EC2-backed cluster means choosing instance types, patching the container instances and tuning placement strategies such as binpack, all of which is undifferentiated work for a team with no Linux administrators.
Option C adds a second concern on top of that one. Spot capacity is the right lever when cost matters more than interruption, but the question says the team accepts a premium, and Spot still leaves them owning the EC2 instances.
Option D discards the orchestrator altogether. Without ECS scheduling there is nothing to replace a failed container, scale with the burst, or register tasks with a load balancer; it is more operational work than option B, not less.
Reference: Architect for AWS Fargate for Amazon ECS