Amazon AWS Certified Solutions Architect Associate SAA-C03 Scalable Application Platforms and Managed Services Practice Test

 

Topic 08 covers scalable application platforms and managed services for the AWS Certified Solutions Architect – Associate certification. These original practice questions apply the verified SAA-C03 objectives to practical decisions and troubleshooting. Select one answer unless a fixed number is requested. For broader preparation, visit the AWS Certified Solutions Architect Associate SAA-C03 Exam Dumps page. Each option includes an explanation of the relevant behavior and scenario constraints.

Question 1

An image upload needs a short stateless transformation that runs only when an object arrives and normally finishes in seconds. Which compute service is the best fit?

1. Use AWS Lambda.
2. Use AWS Batch.
3. Run the container service on Amazon ECS with AWS Fargate.
4. Run the workload on Amazon EC2.
5. Use Amazon EKS.

Correct Answer: A

 

Correct Answer

Answer A is correct because Lambda runs event-driven code without provisioning servers and is suited to short stateless tasks. This directly satisfies the decisive requirement in the scenario: no server management.

Incorrect Answers

Answer B is incorrect because AWS Batch queues and schedules batch jobs across managed compute environments for large parallel processing workloads. It can be useful in other designs, but it does not satisfy the decisive requirement here: no server management.

Answer C is incorrect because ECS provides container orchestration and Fargate supplies managed compute capacity without customer-managed hosts. It can be useful in other designs, but it does not satisfy the decisive requirement here: no server management.

Answer D is incorrect because EC2 provides virtual machines and the broadest guest OS and instance-level control. It can be useful in other designs, but it does not satisfy the decisive requirement here: no server management.

Answer E is incorrect because EKS provides a managed Kubernetes control plane for workloads that require Kubernetes APIs and ecosystem tooling. It can be useful in other designs, but it does not satisfy the decisive requirement here: no server management.

 

Question 2

A containerized API must run continuously, scale horizontally, and the team does not want to manage worker hosts. Which platform is most appropriate?

1. Use AWS Lambda.
2. Run the workload on Amazon EC2.
3. Use Amazon EKS.
4. Use AWS Batch.
5. Run the container service on Amazon ECS with AWS Fargate.

Correct Answer: E

 

Correct Answer

Answer E is correct because ECS provides container orchestration and Fargate supplies managed compute capacity without customer-managed hosts. This directly satisfies the decisive requirement in the scenario: serverless containers.

Incorrect Answers

Answer A is incorrect because Lambda runs event-driven code without provisioning servers and is suited to short stateless tasks. It can be useful in other designs, but it does not satisfy the decisive requirement here: serverless containers.

Answer B is incorrect because EC2 provides virtual machines and the broadest guest OS and instance-level control. It can be useful in other designs, but it does not satisfy the decisive requirement here: serverless containers.

Answer C is incorrect because EKS provides a managed Kubernetes control plane for workloads that require Kubernetes APIs and ecosystem tooling. It can be useful in other designs, but it does not satisfy the decisive requirement here: serverless containers.

Answer D is incorrect because AWS Batch queues and schedules batch jobs across managed compute environments for large parallel processing workloads. It can be useful in other designs, but it does not satisfy the decisive requirement here: serverless containers.

 

Question 3

A company standardizes on Kubernetes APIs and operators and needs AWS-managed Kubernetes control planes. Which service should be used?

1. Use AWS Lambda.
2. Use AWS Batch.
3. Run the container service on Amazon ECS with AWS Fargate.
4. Run the workload on Amazon EC2.
5. Use Amazon EKS.

Correct Answer: E

 

Correct Answer

Answer E is correct because EKS provides a managed Kubernetes control plane for workloads that require Kubernetes APIs and ecosystem tooling. This directly satisfies the decisive requirement in the scenario: EKS.

Incorrect Answers

Answer A is incorrect because Lambda runs event-driven code without provisioning servers and is suited to short stateless tasks. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

Answer B is incorrect because AWS Batch queues and schedules batch jobs across managed compute environments for large parallel processing workloads. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

Answer C is incorrect because ECS provides container orchestration and Fargate supplies managed compute capacity without customer-managed hosts. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

Answer D is incorrect because EC2 provides virtual machines and the broadest guest OS and instance-level control. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

 

Question 4

A commercial network appliance requires a supported AMI, privileged OS configuration, and kernel-level access. Which compute option fits best?

1. Use Amazon EKS.
2. Run the workload on Amazon EC2.
3. Use AWS Batch.
4. Run the container service on Amazon ECS with AWS Fargate.
5. Use AWS Lambda.

Correct Answer: B

 

Correct Answer

Answer B is correct because EC2 provides virtual machines and the broadest guest OS and instance-level control. This directly satisfies the decisive requirement in the scenario: EC2 instances.

Incorrect Answers

Answer A is incorrect because EKS provides a managed Kubernetes control plane for workloads that require Kubernetes APIs and ecosystem tooling. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2 instances.

Answer C is incorrect because AWS Batch queues and schedules batch jobs across managed compute environments for large parallel processing workloads. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2 instances.

Answer D is incorrect because ECS provides container orchestration and Fargate supplies managed compute capacity without customer-managed hosts. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2 instances.

Answer E is incorrect because Lambda runs event-driven code without provisioning servers and is suited to short stateless tasks. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2 instances.

 

Question 5

A nightly genomics job consists of thousands of independent container jobs that can be queued and scheduled when capacity is available. Which managed service is best suited?

1. Run the container service on Amazon ECS with AWS Fargate.
2. Use AWS Lambda.
3. Run the workload on Amazon EC2.
4. Use Amazon EKS.
5. Use AWS Batch.

Correct Answer: E

 

Correct Answer

Answer E is correct because AWS Batch queues and schedules batch jobs across managed compute environments for large parallel processing workloads. This directly satisfies the decisive requirement in the scenario: AWS Batch.

Incorrect Answers

Answer A is incorrect because ECS provides container orchestration and Fargate supplies managed compute capacity without customer-managed hosts. It can be useful in other designs, but it does not satisfy the decisive requirement here: AWS Batch.

Answer B is incorrect because Lambda runs event-driven code without provisioning servers and is suited to short stateless tasks. It can be useful in other designs, but it does not satisfy the decisive requirement here: AWS Batch.

Answer C is incorrect because EC2 provides virtual machines and the broadest guest OS and instance-level control. It can be useful in other designs, but it does not satisfy the decisive requirement here: AWS Batch.

Answer D is incorrect because EKS provides a managed Kubernetes control plane for workloads that require Kubernetes APIs and ecosystem tooling. It can be useful in other designs, but it does not satisfy the decisive requirement here: AWS Batch.

 

Question 6

A web tier runs on replaceable instances behind a load balancer. User sessions must survive instance termination. Where should session state be kept?

1. Move business state out of instance-local storage into an appropriate durable managed data store.
2. Store durable uploaded objects in Amazon S3.
3. Store session or hot ephemeral state in Amazon ElastiCache.
4. Add Amazon ElastiCache in front of the database for repeatedly requested data.
5. Use Amazon EFS for the shared Linux file system.

Correct Answer: C

 

Correct Answer

Answer C is correct because ElastiCache provides a managed in-memory data store that can be shared by a horizontally scaled application tier. This directly satisfies the decisive requirement in the scenario: external low-latency session store.

Incorrect Answers

Answer A is incorrect because Replaceable compute must not hold the only copy of durable application state when instance termination is expected. It can be useful in other designs, but it does not satisfy the decisive requirement here: external low-latency session store.

Answer B is incorrect because S3 provides durable object storage independent of the lifecycle of any application instance. It can be useful in other designs, but it does not satisfy the decisive requirement here: external low-latency session store.

Answer D is incorrect because A cache can serve hot read results without re-running the same database query for every request. It can be useful in other designs, but it does not satisfy the decisive requirement here: external low-latency session store.

Answer E is incorrect because EFS provides a managed NFS file system that can be concurrently mounted by multiple compute instances. It can be useful in other designs, but it does not satisfy the decisive requirement here: external low-latency session store.

 

Question 7

Users upload documents to an autoscaled web fleet. Any instance must be able to serve any uploaded document after a scale event. Where should the files be stored?

1. Move business state out of instance-local storage into an appropriate durable managed data store.
2. Store session or hot ephemeral state in Amazon ElastiCache.
3. Add Amazon ElastiCache in front of the database for repeatedly requested data.
4. Use Amazon EFS for the shared Linux file system.
5. Store durable uploaded objects in Amazon S3.

Correct Answer: E

 

Correct Answer

Answer E is correct because S3 provides durable object storage independent of the lifecycle of any application instance. This directly satisfies the decisive requirement in the scenario: object storage.

Incorrect Answers

Answer A is incorrect because Replaceable compute must not hold the only copy of durable application state when instance termination is expected. It can be useful in other designs, but it does not satisfy the decisive requirement here: object storage.

Answer B is incorrect because ElastiCache provides a managed in-memory data store that can be shared by a horizontally scaled application tier. It can be useful in other designs, but it does not satisfy the decisive requirement here: object storage.

Answer C is incorrect because A cache can serve hot read results without re-running the same database query for every request. It can be useful in other designs, but it does not satisfy the decisive requirement here: object storage.

Answer D is incorrect because EFS provides a managed NFS file system that can be concurrently mounted by multiple compute instances. It can be useful in other designs, but it does not satisfy the decisive requirement here: object storage.

 

Question 8

A read-heavy catalog performs the same expensive database query thousands of times per minute, and slight staleness is acceptable. What should be added?

1. Add Amazon ElastiCache in front of the database for repeatedly requested data.
2. Move business state out of instance-local storage into an appropriate durable managed data store.
3. Store durable uploaded objects in Amazon S3.
4. Store session or hot ephemeral state in Amazon ElastiCache.
5. Use Amazon EFS for the shared Linux file system.

Correct Answer: A

 

Correct Answer

Answer A is correct because A cache can serve hot read results without re-running the same database query for every request. This directly satisfies the decisive requirement in the scenario: managed cache.

Incorrect Answers

Answer B is incorrect because Replaceable compute must not hold the only copy of durable application state when instance termination is expected. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed cache.

Answer C is incorrect because S3 provides durable object storage independent of the lifecycle of any application instance. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed cache.

Answer D is incorrect because ElastiCache provides a managed in-memory data store that can be shared by a horizontally scaled application tier. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed cache.

Answer E is incorrect because EFS provides a managed NFS file system that can be concurrently mounted by multiple compute instances. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed cache.

 

Question 9

A legacy Linux application expects a shared POSIX file system mounted by many instances. Which storage service best externalizes that state?

3. Store durable uploaded objects in Amazon S3.
4. Store session or hot ephemeral state in Amazon ElastiCache.
5. Move business state out of instance-local storage into an appropriate durable managed data store.
6. Add Amazon ElastiCache in front of the database for repeatedly requested data.
7. Use Amazon EFS for the shared Linux file system.

Correct Answer: E

 

Correct Answer

Answer E is correct because EFS provides a managed NFS file system that can be concurrently mounted by multiple compute instances. This directly satisfies the decisive requirement in the scenario: NFS semantics.

Incorrect Answers

Answer A is incorrect because S3 provides durable object storage independent of the lifecycle of any application instance. It can be useful in other designs, but it does not satisfy the decisive requirement here: NFS semantics.

Answer B is incorrect because ElastiCache provides a managed in-memory data store that can be shared by a horizontally scaled application tier. It can be useful in other designs, but it does not satisfy the decisive requirement here: NFS semantics.

Answer C is incorrect because Replaceable compute must not hold the only copy of durable application state when instance termination is expected. It can be useful in other designs, but it does not satisfy the decisive requirement here: NFS semantics.

Answer D is incorrect because A cache can serve hot read results without re-running the same database query for every request. It can be useful in other designs, but it does not satisfy the decisive requirement here: NFS semantics.

 

Question 10

A stateless API currently writes temporary business records to instance store, causing data loss when instances are replaced. Which architectural change is required?

3. Store durable uploaded objects in Amazon S3.
4. Add Amazon ElastiCache in front of the database for repeatedly requested data.
5. Store session or hot ephemeral state in Amazon ElastiCache.
6. Move business state out of instance-local storage into an appropriate durable managed data store.
7. Use Amazon EFS for the shared Linux file system.

Correct Answer: D

 

Correct Answer

Answer D is correct because Replaceable compute must not hold the only copy of durable application state when instance termination is expected. This directly satisfies the decisive requirement in the scenario: move durable state to managed persistent service.

Incorrect Answers

Answer A is incorrect because S3 provides durable object storage independent of the lifecycle of any application instance. It can be useful in other designs, but it does not satisfy the decisive requirement here: move durable state to managed persistent service.

Answer B is incorrect because A cache can serve hot read results without re-running the same database query for every request. It can be useful in other designs, but it does not satisfy the decisive requirement here: move durable state to managed persistent service.

Answer C is incorrect because ElastiCache provides a managed in-memory data store that can be shared by a horizontally scaled application tier. It can be useful in other designs, but it does not satisfy the decisive requirement here: move durable state to managed persistent service.

Answer E is incorrect because EFS provides a managed NFS file system that can be concurrently mounted by multiple compute instances. It can be useful in other designs, but it does not satisfy the decisive requirement here: move durable state to managed persistent service.

 

Question 11

Traffic to the HTTP frontend doubles, but background workers remain lightly loaded. Which architecture supports scaling the web tier without scaling workers?

1. Run web and worker components as separate scalable services or Auto Scaling groups.
2. Use a shared managed file service such as Amazon EFS when POSIX sharing is required.
3. Add database read replicas or another read-scaling mechanism appropriate to the engine.
4. Use API Gateway at the ingress and decouple long work through an asynchronous backend such as SQS.
5. Scale the worker service using queue depth or backlog age.

Correct Answer: A

 

Correct Answer

Answer A is correct because Independent services let each tier scale on its own demand signal instead of forcing one capacity decision across the application. This directly satisfies the decisive requirement in the scenario: separate scalable services.

Incorrect Answers

Answer B is incorrect because A shared file service removes duplicated local copies and lets the compute fleet scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: separate scalable services.

Answer C is incorrect because Read replicas can increase read capacity while leaving the writer focused on transactional writes. It can be useful in other designs, but it does not satisfy the decisive requirement here: separate scalable services.

Answer D is incorrect because Managed ingress and a queue separate request acceptance from worker capacity so the two tiers can scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: separate scalable services.

Answer E is incorrect because Queue backlog is a direct demand signal for workers and need not scale the web tier at the same time. It can be useful in other designs, but it does not satisfy the decisive requirement here: separate scalable services.

 

Question 12

A producer creates queue messages faster than workers can drain them, while the web tier is healthy. What should scale based on backlog?

1. Add database read replicas or another read-scaling mechanism appropriate to the engine.
2. Use a shared managed file service such as Amazon EFS when POSIX sharing is required.
3. Scale the worker service using queue depth or backlog age.
4. Use API Gateway at the ingress and decouple long work through an asynchronous backend such as SQS.
5. Run web and worker components as separate scalable services or Auto Scaling groups.

Correct Answer: C

 

Correct Answer

Answer C is correct because Queue backlog is a direct demand signal for workers and need not scale the web tier at the same time. This directly satisfies the decisive requirement in the scenario: worker fleet.

Incorrect Answers

Answer A is incorrect because Read replicas can increase read capacity while leaving the writer focused on transactional writes. It can be useful in other designs, but it does not satisfy the decisive requirement here: worker fleet.

Answer B is incorrect because A shared file service removes duplicated local copies and lets the compute fleet scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: worker fleet.

Answer D is incorrect because Managed ingress and a queue separate request acceptance from worker capacity so the two tiers can scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: worker fleet.

Answer E is incorrect because Independent services let each tier scale on its own demand signal instead of forcing one capacity decision across the application. It can be useful in other designs, but it does not satisfy the decisive requirement here: worker fleet.

 

Question 13

Read traffic has grown while writes remain stable. Which database-side scaling approach best isolates read demand?

1. Use a shared managed file service such as Amazon EFS when POSIX sharing is required.
2. Add database read replicas or another read-scaling mechanism appropriate to the engine.
3. Scale the worker service using queue depth or backlog age.
4. Use API Gateway at the ingress and decouple long work through an asynchronous backend such as SQS.
5. Run web and worker components as separate scalable services or Auto Scaling groups.

Correct Answer: B

 

Correct Answer

Answer B is correct because Read replicas can increase read capacity while leaving the writer focused on transactional writes. This directly satisfies the decisive requirement in the scenario: add read capacity rather than writer capacity.

Incorrect Answers

Answer A is incorrect because A shared file service removes duplicated local copies and lets the compute fleet scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: add read capacity rather than writer capacity.

Answer C is incorrect because Queue backlog is a direct demand signal for workers and need not scale the web tier at the same time. It can be useful in other designs, but it does not satisfy the decisive requirement here: add read capacity rather than writer capacity.

Answer D is incorrect because Managed ingress and a queue separate request acceptance from worker capacity so the two tiers can scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: add read capacity rather than writer capacity.

Answer E is incorrect because Independent services let each tier scale on its own demand signal instead of forcing one capacity decision across the application. It can be useful in other designs, but it does not satisfy the decisive requirement here: add read capacity rather than writer capacity.

 

Question 14

Hundreds of application instances need the same shared file tree. Which design prevents every web instance from carrying its own copy?

1. Run web and worker components as separate scalable services or Auto Scaling groups.
2. Scale the worker service using queue depth or backlog age.
3. Use API Gateway at the ingress and decouple long work through an asynchronous backend such as SQS.
4. Add database read replicas or another read-scaling mechanism appropriate to the engine.
5. Use a shared managed file service such as Amazon EFS when POSIX sharing is required.

Correct Answer: E

 

Correct Answer

Answer E is correct because A shared file service removes duplicated local copies and lets the compute fleet scale independently. This directly satisfies the decisive requirement in the scenario: external shared file service.

Incorrect Answers

Answer A is incorrect because Independent services let each tier scale on its own demand signal instead of forcing one capacity decision across the application. It can be useful in other designs, but it does not satisfy the decisive requirement here: external shared file service.

Answer B is incorrect because Queue backlog is a direct demand signal for workers and need not scale the web tier at the same time. It can be useful in other designs, but it does not satisfy the decisive requirement here: external shared file service.

Answer C is incorrect because Managed ingress and a queue separate request acceptance from worker capacity so the two tiers can scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: external shared file service.

Answer D is incorrect because Read replicas can increase read capacity while leaving the writer focused on transactional writes. It can be useful in other designs, but it does not satisfy the decisive requirement here: external shared file service.

 

Question 15

A bursty public API should scale its request ingress separately from its compute workers. Which boundary best supports this?

1. Use a shared managed file service such as Amazon EFS when POSIX sharing is required.
2. Use API Gateway at the ingress and decouple long work through an asynchronous backend such as SQS.
3. Add database read replicas or another read-scaling mechanism appropriate to the engine.
4. Run web and worker components as separate scalable services or Auto Scaling groups.
5. Scale the worker service using queue depth or backlog age.

Correct Answer: B

 

Correct Answer

Answer B is correct because Managed ingress and a queue separate request acceptance from worker capacity so the two tiers can scale independently. This directly satisfies the decisive requirement in the scenario: managed API front door plus asynchronous backend.

Incorrect Answers

Answer A is incorrect because A shared file service removes duplicated local copies and lets the compute fleet scale independently. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed API front door plus asynchronous backend.

Answer C is incorrect because Read replicas can increase read capacity while leaving the writer focused on transactional writes. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed API front door plus asynchronous backend.

Answer D is incorrect because Independent services let each tier scale on its own demand signal instead of forcing one capacity decision across the application. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed API front door plus asynchronous backend.

Answer E is incorrect because Queue backlog is a direct demand signal for workers and need not scale the web tier at the same time. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed API front door plus asynchronous backend.

 

Question 16

A team wants AWS-native orchestration for a continuously running container service and does not require Kubernetes compatibility. Which service should coordinate the containers?

1. Use an Amazon ECS service.
2. Use AWS Fargate capacity.
3. Use EC2-backed container capacity.
4. Use AWS Step Functions to own the business workflow state.
5. Use Amazon EKS.

Correct Answer: A

 

Correct Answer

Answer A is correct because ECS is AWS-native container orchestration for long-running services and tasks without requiring Kubernetes semantics. This directly satisfies the decisive requirement in the scenario: ECS service.

Incorrect Answers

Answer B is incorrect because Fargate runs containers without the customer managing EC2 worker nodes. It can be useful in other designs, but it does not satisfy the decisive requirement here: ECS service.

Answer C is incorrect because EC2 capacity preserves control over host instance type, host configuration, and node-level agents. It can be useful in other designs, but it does not satisfy the decisive requirement here: ECS service.

Answer D is incorrect because Step Functions coordinates tasks, waits, retries, branching, and state across distributed work rather than using the container scheduler as a workflow engine. It can be useful in other designs, but it does not satisfy the decisive requirement here: ECS service.

Answer E is incorrect because EKS is the managed AWS service for workloads that require the Kubernetes API and ecosystem. It can be useful in other designs, but it does not satisfy the decisive requirement here: ECS service.

 

Question 17

An acquired platform depends on Kubernetes CRDs, Helm charts, and operators. Which orchestration service preserves those interfaces?

1. Use AWS Fargate capacity.
2. Use Amazon EKS.
3. Use AWS Step Functions to own the business workflow state.
4. Use EC2-backed container capacity.
5. Use an Amazon ECS service.

Correct Answer: B

 

Correct Answer

Answer B is correct because EKS is the managed AWS service for workloads that require the Kubernetes API and ecosystem. This directly satisfies the decisive requirement in the scenario: EKS.

Incorrect Answers

Answer A is incorrect because Fargate runs containers without the customer managing EC2 worker nodes. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

Answer C is incorrect because Step Functions coordinates tasks, waits, retries, branching, and state across distributed work rather than using the container scheduler as a workflow engine. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

Answer D is incorrect because EC2 capacity preserves control over host instance type, host configuration, and node-level agents. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

Answer E is incorrect because ECS is AWS-native container orchestration for long-running services and tasks without requiring Kubernetes semantics. It can be useful in other designs, but it does not satisfy the decisive requirement here: EKS.

 

Question 18

A small container workload should run without provisioning or patching EC2 worker nodes. Which compute mode should back the orchestrator?

1. Use an Amazon ECS service.
2. Use EC2-backed container capacity.
3. Use AWS Step Functions to own the business workflow state.
4. Use Amazon EKS.
5. Use AWS Fargate capacity.

Correct Answer: E

 

Correct Answer

Answer E is correct because Fargate runs containers without the customer managing EC2 worker nodes. This directly satisfies the decisive requirement in the scenario: Fargate.

Incorrect Answers

Answer A is incorrect because ECS is AWS-native container orchestration for long-running services and tasks without requiring Kubernetes semantics. It can be useful in other designs, but it does not satisfy the decisive requirement here: Fargate.

Answer B is incorrect because EC2 capacity preserves control over host instance type, host configuration, and node-level agents. It can be useful in other designs, but it does not satisfy the decisive requirement here: Fargate.

Answer C is incorrect because Step Functions coordinates tasks, waits, retries, branching, and state across distributed work rather than using the container scheduler as a workflow engine. It can be useful in other designs, but it does not satisfy the decisive requirement here: Fargate.

Answer D is incorrect because EKS is the managed AWS service for workloads that require the Kubernetes API and ecosystem. It can be useful in other designs, but it does not satisfy the decisive requirement here: Fargate.

 

Question 19

A legacy service requires specific EC2 instance families, local agents, and privileged host configuration. Which container capacity model is most appropriate?

1. Use Amazon EKS.
2. Use AWS Fargate capacity.
3. Use an Amazon ECS service.
4. Use EC2-backed container capacity.
5. Use AWS Step Functions to own the business workflow state.

Correct Answer: D

 

Correct Answer

Answer D is correct because EC2 capacity preserves control over host instance type, host configuration, and node-level agents. This directly satisfies the decisive requirement in the scenario: EC2-backed capacity.

Incorrect Answers

Answer A is incorrect because EKS is the managed AWS service for workloads that require the Kubernetes API and ecosystem. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2-backed capacity.

Answer B is incorrect because Fargate runs containers without the customer managing EC2 worker nodes. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2-backed capacity.

Answer C is incorrect because ECS is AWS-native container orchestration for long-running services and tasks without requiring Kubernetes semantics. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2-backed capacity.

Answer E is incorrect because Step Functions coordinates tasks, waits, retries, branching, and state across distributed work rather than using the container scheduler as a workflow engine. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2-backed capacity.

 

Question 20

A distributed business process coordinates container tasks, waits for external approvals, and retries failed steps. Which service should own the workflow state?

1. Use an Amazon ECS service.
2. Use Amazon EKS.
3. Use AWS Step Functions to own the business workflow state.
4. Use AWS Fargate capacity.
5. Use EC2-backed container capacity.

Correct Answer: C

 

Correct Answer

Answer C is correct because Step Functions coordinates tasks, waits, retries, branching, and state across distributed work rather than using the container scheduler as a workflow engine. This directly satisfies the decisive requirement in the scenario: Step Functions.

Incorrect Answers

Answer A is incorrect because ECS is AWS-native container orchestration for long-running services and tasks without requiring Kubernetes semantics. It can be useful in other designs, but it does not satisfy the decisive requirement here: Step Functions.

Answer B is incorrect because EKS is the managed AWS service for workloads that require the Kubernetes API and ecosystem. It can be useful in other designs, but it does not satisfy the decisive requirement here: Step Functions.

Answer D is incorrect because Fargate runs containers without the customer managing EC2 worker nodes. It can be useful in other designs, but it does not satisfy the decisive requirement here: Step Functions.

Answer E is incorrect because EC2 capacity preserves control over host instance type, host configuration, and node-level agents. It can be useful in other designs, but it does not satisfy the decisive requirement here: Step Functions.

 

Question 21

A support application must extract sentiment and key phrases from customer messages without training a custom NLP platform. Which managed service is the best fit?

1. Use Amazon Polly.
2. Use Amazon OpenSearch Service.
3. Use AWS Transfer Family.
4. Use Amazon Comprehend.
5. Use Amazon Rekognition.

Correct Answer: D

 

Correct Answer

Answer D is correct because Amazon Comprehend provides managed natural-language processing capabilities such as sentiment and entity analysis. This directly satisfies the decisive requirement in the scenario: managed NLP.

Incorrect Answers

Answer A is incorrect because Amazon Polly converts text into lifelike speech through a managed text-to-speech service. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed NLP.

Answer B is incorrect because OpenSearch Service provides managed search and analytics over indexed documents and logs. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed NLP.

Answer C is incorrect because Transfer Family provides managed file-transfer protocol endpoints such as SFTP with AWS storage backends. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed NLP.

Answer E is incorrect because Amazon Rekognition provides managed image and video analysis capabilities including labels and object detection. It can be useful in other designs, but it does not satisfy the decisive requirement here: managed NLP.

 

Question 22

An accessibility feature must synthesize natural speech from text in multiple voices. Which managed service should be used?

1. Use Amazon OpenSearch Service.
2. Use Amazon Comprehend.
3. Use AWS Transfer Family.
4. Use Amazon Polly.
5. Use Amazon Rekognition.

Correct Answer: D

 

Correct Answer

Answer D is correct because Amazon Polly converts text into lifelike speech through a managed text-to-speech service. This directly satisfies the decisive requirement in the scenario: Amazon Polly.

Incorrect Answers

Answer A is incorrect because OpenSearch Service provides managed search and analytics over indexed documents and logs. It can be useful in other designs, but it does not satisfy the decisive requirement here: Amazon Polly.

Answer B is incorrect because Amazon Comprehend provides managed natural-language processing capabilities such as sentiment and entity analysis. It can be useful in other designs, but it does not satisfy the decisive requirement here: Amazon Polly.

Answer C is incorrect because Transfer Family provides managed file-transfer protocol endpoints such as SFTP with AWS storage backends. It can be useful in other designs, but it does not satisfy the decisive requirement here: Amazon Polly.

Answer E is incorrect because Amazon Rekognition provides managed image and video analysis capabilities including labels and object detection. It can be useful in other designs, but it does not satisfy the decisive requirement here: Amazon Polly.

 

Question 23

Partners must upload files over SFTP into Amazon S3 without the company operating SFTP servers. Which service best fits?

1. Use Amazon Comprehend.
2. Use Amazon Polly.
3. Use Amazon Rekognition.
4. Use Amazon OpenSearch Service.
5. Use AWS Transfer Family.

Correct Answer: E

 

Correct Answer

Answer E is correct because Transfer Family provides managed file-transfer protocol endpoints such as SFTP with AWS storage backends. This directly satisfies the decisive requirement in the scenario: Transfer Family.

Incorrect Answers

Answer A is incorrect because Amazon Comprehend provides managed natural-language processing capabilities such as sentiment and entity analysis. It can be useful in other designs, but it does not satisfy the decisive requirement here: Transfer Family.

Answer B is incorrect because Amazon Polly converts text into lifelike speech through a managed text-to-speech service. It can be useful in other designs, but it does not satisfy the decisive requirement here: Transfer Family.

Answer C is incorrect because Amazon Rekognition provides managed image and video analysis capabilities including labels and object detection. It can be useful in other designs, but it does not satisfy the decisive requirement here: Transfer Family.

Answer D is incorrect because OpenSearch Service provides managed search and analytics over indexed documents and logs. It can be useful in other designs, but it does not satisfy the decisive requirement here: Transfer Family.

 

Question 24

An application needs full-text indexing and low-latency search over log and product documents. Which managed service is appropriate?

1. Use Amazon Polly.
2. Use Amazon Comprehend.
3. Use AWS Transfer Family.
4. Use Amazon OpenSearch Service.
5. Use Amazon Rekognition.

Correct Answer: D

 

Correct Answer

Answer D is correct because OpenSearch Service provides managed search and analytics over indexed documents and logs. This directly satisfies the decisive requirement in the scenario: OpenSearch Service.

Incorrect Answers

Answer A is incorrect because Amazon Polly converts text into lifelike speech through a managed text-to-speech service. It can be useful in other designs, but it does not satisfy the decisive requirement here: OpenSearch Service.

Answer B is incorrect because Amazon Comprehend provides managed natural-language processing capabilities such as sentiment and entity analysis. It can be useful in other designs, but it does not satisfy the decisive requirement here: OpenSearch Service.

Answer C is incorrect because Transfer Family provides managed file-transfer protocol endpoints such as SFTP with AWS storage backends. It can be useful in other designs, but it does not satisfy the decisive requirement here: OpenSearch Service.

Answer E is incorrect because Amazon Rekognition provides managed image and video analysis capabilities including labels and object detection. It can be useful in other designs, but it does not satisfy the decisive requirement here: OpenSearch Service.

 

Question 25

A workflow must detect labels and objects in uploaded images without building its own computer-vision model. Which managed service should be used?

1. Use Amazon Comprehend.
2. Use AWS Transfer Family.
3. Use Amazon Polly.
4. Use Amazon Rekognition.
5. Use Amazon OpenSearch Service.

Correct Answer: D

 

Correct Answer

Answer D is correct because Amazon Rekognition provides managed image and video analysis capabilities including labels and object detection. This directly satisfies the decisive requirement in the scenario: Rekognition.

Incorrect Answers

Answer A is incorrect because Amazon Comprehend provides managed natural-language processing capabilities such as sentiment and entity analysis. It can be useful in other designs, but it does not satisfy the decisive requirement here: Rekognition.

Answer B is incorrect because Transfer Family provides managed file-transfer protocol endpoints such as SFTP with AWS storage backends. It can be useful in other designs, but it does not satisfy the decisive requirement here: Rekognition.

Answer C is incorrect because Amazon Polly converts text into lifelike speech through a managed text-to-speech service. It can be useful in other designs, but it does not satisfy the decisive requirement here: Rekognition.

Answer E is incorrect because OpenSearch Service provides managed search and analytics over indexed documents and logs. It can be useful in other designs, but it does not satisfy the decisive requirement here: Rekognition.

 

Question 26

A Java service already runs in containers on premises. The team wants to move it first without rewriting application logic. Which first step best preserves portability?

1. Adopt immutable deployment using versioned machine or container images and replacement rather than in-place drift.
2. Use path- or host-based routing at an application ingress to send selected capabilities to new services.
3. Move the containerized service to Amazon ECS or EKS using the existing image contract.
4. Run the application on Amazon EC2.
5. Use Amazon CloudFront for cacheable static content.

Correct Answer: C

 

Correct Answer

Answer C is correct because Container orchestration preserves the existing packaging boundary while moving runtime operations to AWS. This directly satisfies the decisive requirement in the scenario: run containers on managed AWS orchestration.

Incorrect Answers

Answer A is incorrect because Immutable replacement makes releases reproducible and reduces configuration differences between long-lived servers. It can be useful in other designs, but it does not satisfy the decisive requirement here: run containers on managed AWS orchestration.

Answer B is incorrect because Strangler routing lets a monolith and extracted services coexist while functionality is migrated incrementally. It can be useful in other designs, but it does not satisfy the decisive requirement here: run containers on managed AWS orchestration.

Answer D is incorrect because EC2 is appropriate when the workload still requires a specific guest OS, license model, or administrative control. It can be useful in other designs, but it does not satisfy the decisive requirement here: run containers on managed AWS orchestration.

Answer E is incorrect because CloudFront can modernize delivery of cacheable assets independently of the origin application implementation. It can be useful in other designs, but it does not satisfy the decisive requirement here: run containers on managed AWS orchestration.

 

Question 27

A commercial application uses a license tied to a supported Windows server and requires administrative access. Which migration target minimizes application change?

2. Run the application on Amazon EC2.
3. Move the containerized service to Amazon ECS or EKS using the existing image contract.
4. Adopt immutable deployment using versioned machine or container images and replacement rather than in-place drift.
5. Use Amazon CloudFront for cacheable static content.
6. Use path- or host-based routing at an application ingress to send selected capabilities to new services.

Correct Answer: A

 

Correct Answer

Answer A is correct because EC2 is appropriate when the workload still requires a specific guest OS, license model, or administrative control. This directly satisfies the decisive requirement in the scenario: EC2.

Incorrect Answers

Answer B is incorrect because Container orchestration preserves the existing packaging boundary while moving runtime operations to AWS. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2.

Answer C is incorrect because Immutable replacement makes releases reproducible and reduces configuration differences between long-lived servers. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2.

Answer D is incorrect because CloudFront can modernize delivery of cacheable assets independently of the origin application implementation. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2.

Answer E is incorrect because Strangler routing lets a monolith and extracted services coexist while functionality is migrated incrementally. It can be useful in other designs, but it does not satisfy the decisive requirement here: EC2.

 

Question 28

A fleet must replace servers consistently from a tested image instead of patching long-lived instances in place. Which deployment style should be adopted?

1. Move the containerized service to Amazon ECS or EKS using the existing image contract.
2. Use Amazon CloudFront for cacheable static content.
3. Adopt immutable deployment using versioned machine or container images and replacement rather than in-place drift.
4. Use path- or host-based routing at an application ingress to send selected capabilities to new services.
5. Run the application on Amazon EC2.

Correct Answer: C

 

Correct Answer

Answer C is correct because Immutable replacement makes releases reproducible and reduces configuration differences between long-lived servers. This directly satisfies the decisive requirement in the scenario: immutable images.

Incorrect Answers

Answer A is incorrect because Container orchestration preserves the existing packaging boundary while moving runtime operations to AWS. It can be useful in other designs, but it does not satisfy the decisive requirement here: immutable images.

Answer B is incorrect because CloudFront can modernize delivery of cacheable assets independently of the origin application implementation. It can be useful in other designs, but it does not satisfy the decisive requirement here: immutable images.

Answer D is incorrect because Strangler routing lets a monolith and extracted services coexist while functionality is migrated incrementally. It can be useful in other designs, but it does not satisfy the decisive requirement here: immutable images.

Answer E is incorrect because EC2 is appropriate when the workload still requires a specific guest OS, license model, or administrative control. It can be useful in other designs, but it does not satisfy the decisive requirement here: immutable images.

 

Question 29

A global website has a mostly static asset tier that can be modernized independently from the application servers. Which service can move delivery closer to users first?

1. Move the containerized service to Amazon ECS or EKS using the existing image contract.
2. Use Amazon CloudFront for cacheable static content.
3. Adopt immutable deployment using versioned machine or container images and replacement rather than in-place drift.
4. Run the application on Amazon EC2.
5. Use path- or host-based routing at an application ingress to send selected capabilities to new services.

Correct Answer: B

 

Correct Answer

Answer B is correct because CloudFront can modernize delivery of cacheable assets independently of the origin application implementation. This directly satisfies the decisive requirement in the scenario: CloudFront edge cache.

Incorrect Answers

Answer A is incorrect because Container orchestration preserves the existing packaging boundary while moving runtime operations to AWS. It can be useful in other designs, but it does not satisfy the decisive requirement here: CloudFront edge cache.

Answer C is incorrect because Immutable replacement makes releases reproducible and reduces configuration differences between long-lived servers. It can be useful in other designs, but it does not satisfy the decisive requirement here: CloudFront edge cache.

Answer D is incorrect because EC2 is appropriate when the workload still requires a specific guest OS, license model, or administrative control. It can be useful in other designs, but it does not satisfy the decisive requirement here: CloudFront edge cache.

Answer E is incorrect because Strangler routing lets a monolith and extracted services coexist while functionality is migrated incrementally. It can be useful in other designs, but it does not satisfy the decisive requirement here: CloudFront edge cache.

 

Question 30

A monolith must be decomposed gradually, routing only selected URL paths to new services while the rest stays unchanged. Which migration pattern best supports this?

1. Adopt immutable deployment using versioned machine or container images and replacement rather than in-place drift.
2. Use Amazon CloudFront for cacheable static content.
3. Move the containerized service to Amazon ECS or EKS using the existing image contract.
4. Use path- or host-based routing at an application ingress to send selected capabilities to new services.
5. Run the application on Amazon EC2.

Correct Answer: D

 

Correct Answer

Answer D is correct because Strangler routing lets a monolith and extracted services coexist while functionality is migrated incrementally. This directly satisfies the decisive requirement in the scenario: strangler routing.

Incorrect Answers

Answer A is incorrect because Immutable replacement makes releases reproducible and reduces configuration differences between long-lived servers. It can be useful in other designs, but it does not satisfy the decisive requirement here: strangler routing.

Answer B is incorrect because CloudFront can modernize delivery of cacheable assets independently of the origin application implementation. It can be useful in other designs, but it does not satisfy the decisive requirement here: strangler routing.

Answer C is incorrect because Container orchestration preserves the existing packaging boundary while moving runtime operations to AWS. It can be useful in other designs, but it does not satisfy the decisive requirement here: strangler routing.

Answer E is incorrect because EC2 is appropriate when the workload still requires a specific guest OS, license model, or administrative control. It can be useful in other designs, but it does not satisfy the decisive requirement here: strangler routing.