A company runs its containers on Amazon ECS and currently manages the cluster of Amazon EC2 instances that host the containers, spending time patching and scaling those hosts. It wants to keep using ECS but stop provisioning, patching and scaling any host servers itself. Which compute option should it select for its ECS tasks?
- AAmazon EC2 launch type, where the company keeps running and patching the pool of host instances that the ECS containers are placed onto
- BAmazon EKS, which would replace ECS with managed Kubernetes but still leaves the company managing its own worker node instances by default
- CAWS Lambda, which runs function code in response to events rather than running the company's existing long-lived ECS container tasks
- DAWS Fargate, the serverless compute engine that runs ECS containers without the company provisioning, patching or scaling any host servers Correct
Why A is wrong: The EC2 launch type still requires the company to run and patch the host instances, which is exactly the management burden it wants to remove.
Why B is wrong: Switching to EKS changes the orchestrator and, without a serverless engine, still leaves worker nodes to manage, so it does not meet the stated goal.
Why C is wrong: Lambda runs event-driven functions, not the company's existing ECS container tasks, so it does not fit a request to keep using ECS.
Why D is correct: Fargate is the serverless engine for ECS that runs each task without the customer managing host instances, removing the host patching and scaling work.