Topic 07 Practice Test 1 covers Enterprise Integration, Event-Driven Patterns, Legacy Systems, and Data Synchronization in Domain 2 (Task 2.3). For broader exam preparation, review the AIP-C01 Exam Dumps. Every option includes focused technical reasoning tied to the tested decision.
Question 1
An operations team is improving the reliability of an FM input path. A mainframe-adjacent application already exposes a fragile HTTP interface, but the new GenAI component needs a stable managed API contract. Which option best meets the operational requirement? Choose ONE.
- Amazon SQS buffering between enterprise systems
- EventBridge event-driven decoupling
- Step Functions enterprise workflow orchestration
- API Gateway facade for a legacy HTTP service
Correct Answer: D
Correct Answer
Answer D is correct because The requirement points to API Gateway facade for a legacy HTTP service. It is meant to present a managed API boundary that shields callers from legacy endpoint details and supports validation, authorization, and versioning. The scenario needs an approach where an existing enterprise application must expose a stable integration contract without rewriting the legacy backend.
Incorrect Answers
Answer A is incorrect because Amazon SQS buffering between enterprise systems is relevant, but it is intended to queue integration work so bursts and temporary downstream outages do not overload the FM consumer when producers can outpace the GenAI worker and messages must survive transient failures. The actual requirement is an existing enterprise application must expose a stable integration contract without rewriting the legacy backend.
Answer B is incorrect because EventBridge event-driven decoupling solves a different input problem. It helps to publish business events to a bus so producers and GenAI consumers evolve independently, normally when the integration must be loosely coupled and react to enterprise events without synchronous dependencies. The current workflow needs an existing enterprise application must expose a stable integration contract without rewriting the legacy backend.
Answer C is incorrect because Step Functions enterprise workflow orchestration is appropriate where a business process spans several services and needs durable orchestration rather than one long-running function. It achieves that by helping to coordinate multiple integration steps with explicit state, retries, branches, and compensating paths. This scenario instead requires an existing enterprise application must expose a stable integration contract without rewriting the legacy backend.
Question 2
A developer is hardening an FM-consumption workflow. Customer account updates in a supported SaaS CRM must be copied into an AWS data store for downstream GenAI use. Which solution most directly resolves this issue? Choose ONE.
- DynamoDB Streams change-event integration
- Amazon SQS buffering between enterprise systems
- EventBridge Pipes filtering and enrichment
- Amazon AppFlow SaaS data synchronization
Correct Answer: D
Correct Answer
Answer D is correct because Amazon AppFlow SaaS data synchronization matches the constraint. It can move selected data between supported SaaS applications and AWS on a managed schedule or event-driven flow. This is the intended approach when enterprise SaaS records must be synchronized into AWS without maintaining custom extraction code.
Incorrect Answers
Answer A is incorrect because DynamoDB Streams change-event integration can help a pipeline emit item-level change records that can trigger downstream processing when application state changes, especially when updates already land in DynamoDB and the GenAI workflow should react only to actual table changes. The present requirement is enterprise SaaS records must be synchronized into AWS without maintaining custom extraction code, so it is not the best answer.
Answer B is incorrect because Amazon SQS buffering between enterprise systems would improve another stage by enabling it to queue integration work so bursts and temporary downstream outages do not overload the FM consumer. It fits when producers can outpace the GenAI worker and messages must survive transient failures. The stated problem is enterprise SaaS records must be synchronized into AWS without maintaining custom extraction code.
Answer C is incorrect because EventBridge Pipes filtering and enrichment works for cases where only a subset of events should reach the GenAI integration and simple routing logic should not require a custom service by helping to connect an event source to a target with declarative filtering, transformation, and optional enrichment. The present scenario needs enterprise SaaS records must be synchronized into AWS without maintaining custom extraction code.
Question 3
A GenAI workload is moving from prototype to production. A completed document classification result must notify billing, compliance, and analytics services without coupling those consumers together. Which design is the best fit for this requirement? Choose ONE.
- Step Functions enterprise workflow orchestration
- Amazon SNS fanout for shared enterprise events
- EventBridge Pipes filtering and enrichment
- EventBridge event-driven decoupling
Correct Answer: B
Correct Answer
Answer B is correct because Use Amazon SNS fanout for shared enterprise events for this design. It will publish one event to multiple subscribed consumers that need independent copies. That aligns with the need that several enterprise applications must react independently to the same GenAI-related event.
Incorrect Answers
Answer A is incorrect because Step Functions enterprise workflow orchestration is useful when a business process spans several services and needs durable orchestration rather than one long-running function and can coordinate multiple integration steps with explicit state, retries, branches, and compensating paths. The described workload requires several enterprise applications must react independently to the same GenAI-related event.
Answer C is incorrect because EventBridge Pipes filtering and enrichment can connect an event source to a target with declarative filtering, transformation, and optional enrichment. It fits when only a subset of events should reach the GenAI integration and simple routing logic should not require a custom service. This scenario instead requires several enterprise applications must react independently to the same GenAI-related event, so it targets a different problem.
Answer D is incorrect because EventBridge event-driven decoupling is useful when the integration must be loosely coupled and react to enterprise events without synchronous dependencies because it can publish business events to a bus so producers and GenAI consumers evolve independently. The current requirement is several enterprise applications must react independently to the same GenAI-related event, which makes this a mismatch.
Question 4
A regulated workflow requires deterministic preprocessing before inference. A SaaS ticketing system sends webhook callbacks when a case changes. Each callback needs signature checking and a small transformation before starting GenAI processing. Which implementation best satisfies the requirement? Choose ONE.
- Amazon SQS buffering between enterprise systems
- DynamoDB Streams change-event integration
- Lambda webhook handler behind API Gateway
- Step Functions enterprise workflow orchestration
Correct Answer: C
Correct Answer
Answer C is correct because Lambda webhook handler behind API Gateway is the best fit because it can accept a short-lived webhook callback, validate it, transform the payload, and invoke downstream AWS services. The workflow needs a case where an external enterprise application pushes lightweight events and the handler should scale without server management.
Incorrect Answers
Answer A is incorrect because Amazon SQS buffering between enterprise systems helps when producers can outpace the GenAI worker and messages must survive transient failures by allowing the pipeline to queue integration work so bursts and temporary downstream outages do not overload the FM consumer. The described issue is an external enterprise application pushes lightweight events and the handler should scale without server management.
Answer B is incorrect because DynamoDB Streams change-event integration is relevant, but it is intended to emit item-level change records that can trigger downstream processing when application state changes when updates already land in DynamoDB and the GenAI workflow should react only to actual table changes. The actual requirement is an external enterprise application pushes lightweight events and the handler should scale without server management.
Answer D is incorrect because Step Functions enterprise workflow orchestration is appropriate where a business process spans several services and needs durable orchestration rather than one long-running function. It achieves that by helping to coordinate multiple integration steps with explicit state, retries, branches, and compensating paths. This scenario instead requires an external enterprise application pushes lightweight events and the handler should scale without server management.
Question 5
During a production-readiness review, engineers identify a data-handling issue. A legal department writes large document batches to an on-premises NFS share and wants them copied to S3 every night for analysis. What should be used to meet this requirement? Choose ONE.
- EventBridge event-driven decoupling
- DynamoDB Streams change-event integration
- AWS DataSync for scheduled on-premises file transfer
- EventBridge Pipes filtering and enrichment
Correct Answer: C
Correct Answer
Answer C is correct because Implement AWS DataSync for scheduled on-premises file transfer. It is designed to transfer large file sets between on-premises storage and AWS with managed scheduling and integrity handling. Here, the important constraint is that enterprise files must be synchronized reliably from on-premises storage before GenAI processing.
Incorrect Answers
Answer A is incorrect because EventBridge event-driven decoupling has value because it can publish business events to a bus so producers and GenAI consumers evolve independently. Its proper fit is when the integration must be loosely coupled and react to enterprise events without synchronous dependencies. This question requires enterprise files must be synchronized reliably from on-premises storage before GenAI processing.
Answer B is incorrect because DynamoDB Streams change-event integration can help a pipeline emit item-level change records that can trigger downstream processing when application state changes, especially when updates already land in DynamoDB and the GenAI workflow should react only to actual table changes. The present requirement is enterprise files must be synchronized reliably from on-premises storage before GenAI processing, so it is not the best answer.
Answer D is incorrect because EventBridge Pipes filtering and enrichment works for cases where only a subset of events should reach the GenAI integration and simple routing logic should not require a custom service by helping to connect an event source to a target with declarative filtering, transformation, and optional enrichment. The present scenario needs enterprise files must be synchronized reliably from on-premises storage before GenAI processing.
Question 6
A cloud engineering team must correct a data-processing weakness. Developers need controlled access to GenAI tooling using the company’s existing identity provider and group memberships. What is the most appropriate implementation choice? Choose ONE.
- IAM permissions boundary for delegated development
- IAM Identity Center federation for workforce access
- STS AssumeRole for cross-account integration
- cross-account resource policy with explicit principals
Correct Answer: B
Correct Answer
Answer B is correct because IAM Identity Center federation for workforce access should be used. It can federate enterprise workforce identities into AWS so access is based on centrally managed users and groups. That capability is needed because employees must use existing corporate identities rather than separate long-lived AWS credentials.
Incorrect Answers
Answer A is incorrect because IAM permissions boundary for delegated development belongs in workflows where teams may self-service roles but central security must prevent them from exceeding an approved privilege ceiling because it can cap the maximum permissions that application teams can grant to roles they create. This question instead asks for employees must use existing corporate identities rather than separate long-lived AWS credentials.
Answer C is incorrect because STS AssumeRole for cross-account integration targets a different issue by helping to issue temporary credentials by assuming a narrowly scoped role in the target account. It fits where an application in one AWS account must call an approved GenAI resource in another account without static access keys. The deciding condition is employees must use existing corporate identities rather than separate long-lived AWS credentials.
Answer D is incorrect because cross-account resource policy with explicit principals can grant a known external account or role access at the resource boundary while keeping the allowed principal set narrow. It fits when a shared resource must be consumed across accounts with auditable, explicit trust. This scenario instead requires employees must use existing corporate identities rather than separate long-lived AWS credentials, so it targets a different problem.
Question 7
A solutions architect is reviewing how source data reaches an FM. A service currently uses an administrator role just to call selected Bedrock models. Security requires the minimum permissions needed for inference. What should the developer implement? Choose ONE.
- IAM permissions boundary for delegated development
- central GenAI gateway with policy enforcement
- AWS PrivateLink private service exposure
- least-privilege IAM policy for Bedrock invocation
Correct Answer: D
Correct Answer
Answer D is correct because least-privilege IAM policy for Bedrock invocation solves the right problem. It will allow only the required model actions and resources instead of broad wildcard permissions. The case depends on the fact that an enterprise application should invoke approved FMs but must not gain unrelated AWS privileges.
Incorrect Answers
Answer A is incorrect because IAM permissions boundary for delegated development can cap the maximum permissions that application teams can grant to roles they create. Its best fit is where teams may self-service roles but central security must prevent them from exceeding an approved privilege ceiling. The current case instead needs an enterprise application should invoke approved FMs but must not gain unrelated AWS privileges.
Answer B is incorrect because central GenAI gateway with policy enforcement helps when many applications need consistent controls without embedding governance logic in every client by allowing the pipeline to route enterprise FM consumption through one abstraction layer that applies authentication, model allowlists, quotas, and logging. The described issue is an enterprise application should invoke approved FMs but must not gain unrelated AWS privileges.
Answer C is incorrect because AWS PrivateLink private service exposure is relevant, but it is intended to publish an internal service through private VPC connectivity without sending traffic over the public internet when enterprise consumers in VPCs need private access to a centralized GenAI integration service. The actual requirement is an enterprise application should invoke approved FMs but must not gain unrelated AWS privileges.
Question 8
A platform team is reviewing a production GenAI data pipeline. A customer-facing app calls a GenAI endpoint through API Gateway and must reject requests without a valid user token. Which implementation provides the required behavior? Choose ONE.
- Cognito JWT authorization at API Gateway
- STS AssumeRole for cross-account integration
- cross-account resource policy with explicit principals
- AWS PrivateLink private service exposure
Correct Answer: A
Correct Answer
Answer A is correct because Cognito JWT authorization at API Gateway is the direct remedy. It can validate application-user tokens at the API boundary before requests enter the GenAI backend. The critical scenario condition is that external application users need authenticated, claim-aware access to a managed GenAI API.
Incorrect Answers
Answer B is incorrect because STS AssumeRole for cross-account integration has value because it can issue temporary credentials by assuming a narrowly scoped role in the target account. Its proper fit is when an application in one AWS account must call an approved GenAI resource in another account without static access keys. This question requires external application users need authenticated, claim-aware access to a managed GenAI API.
Answer C is incorrect because cross-account resource policy with explicit principals can help a pipeline grant a known external account or role access at the resource boundary while keeping the allowed principal set narrow, especially when a shared resource must be consumed across accounts with auditable, explicit trust. The present requirement is external application users need authenticated, claim-aware access to a managed GenAI API, so it is not the best answer.
Answer D is incorrect because AWS PrivateLink private service exposure would improve another stage by enabling it to publish an internal service through private VPC connectivity without sending traffic over the public internet. It fits when enterprise consumers in VPCs need private access to a centralized GenAI integration service. The stated problem is external application users need authenticated, claim-aware access to a managed GenAI API.
Question 9
An enterprise team is refining the input pipeline for a GenAI application. Hundreds of workloads should access only GenAI resources tagged for their department and environment. Which change should the team make? Choose ONE.
- IAM permissions boundary for delegated development
- cross-account resource policy with explicit principals
- central GenAI gateway with policy enforcement
- attribute-based IAM access control
Correct Answer: D
Correct Answer
Answer D is correct because Use attribute-based IAM access control. It will use principal and resource tags in policy conditions so permissions scale with business attributes. The deciding condition is that many teams need access determined by attributes such as department or environment rather than one policy per user. This matches the requirement.
Incorrect Answers
Answer A is incorrect because IAM permissions boundary for delegated development can cap the maximum permissions that application teams can grant to roles they create, which matters when teams may self-service roles but central security must prevent them from exceeding an approved privilege ceiling. The present case is driven by many teams need access determined by attributes such as department or environment rather than one policy per user.
Answer B is incorrect because cross-account resource policy with explicit principals belongs in workflows where a shared resource must be consumed across accounts with auditable, explicit trust because it can grant a known external account or role access at the resource boundary while keeping the allowed principal set narrow. This question instead asks for many teams need access determined by attributes such as department or environment rather than one policy per user.
Answer C is incorrect because central GenAI gateway with policy enforcement is useful when many applications need consistent controls without embedding governance logic in every client and can route enterprise FM consumption through one abstraction layer that applies authentication, model allowlists, quotas, and logging. The described workload requires many teams need access determined by attributes such as department or environment rather than one policy per user.
Question 10
An application team is standardizing data before foundation-model inference. Finance and HR applications call the same GenAI service, yet each should have separate permissions and clearly attributable CloudTrail activity. Which approach most directly addresses the stated constraint? Choose ONE.
- AWS PrivateLink private service exposure
- per-application IAM role isolation
- STS AssumeRole for cross-account integration
- central GenAI gateway with policy enforcement
Correct Answer: B
Correct Answer
Answer B is correct because per-application IAM role isolation directly fits. It can assign separate execution roles to integrating applications so permissions and audit trails are not shared. The scenario is specifically about the case where multiple enterprise apps use the same gateway but each requires distinct least-privilege access and attribution.
Incorrect Answers
Answer A is incorrect because AWS PrivateLink private service exposure is designed to publish an internal service through private VPC connectivity without sending traffic over the public internet. That makes sense when enterprise consumers in VPCs need private access to a centralized GenAI integration service. The question is centered on multiple enterprise apps use the same gateway but each requires distinct least-privilege access and attribution.
Answer C is incorrect because STS AssumeRole for cross-account integration helps when an application in one AWS account must call an approved GenAI resource in another account without static access keys by allowing the pipeline to issue temporary credentials by assuming a narrowly scoped role in the target account. The described issue is multiple enterprise apps use the same gateway but each requires distinct least-privilege access and attribution.
Answer D is incorrect because central GenAI gateway with policy enforcement is relevant, but it is intended to route enterprise FM consumption through one abstraction layer that applies authentication, model allowlists, quotas, and logging when many applications need consistent controls without embedding governance logic in every client. The actual requirement is multiple enterprise apps use the same gateway but each requires distinct least-privilege access and attribution.
Question 11
An operations team is improving the reliability of an FM input path. A hospital must process patient notes inside its own data center and cannot send the raw records to a public Region. Which option best meets the operational requirement? Choose ONE.
- AWS Outposts for local data processing
- AWS Wavelength for edge-proximate inference integration
- store-and-forward integration during link outages
- region-aware request routing for data residency
Correct Answer: A
Correct Answer
Answer A is correct because The correct choice is AWS Outposts for local data processing. It helps to run AWS infrastructure and application components on premises so regulated data can remain at the local site. The deciding detail is that jurisdiction or policy requires sensitive data processing to stay on premises while using AWS-consistent infrastructure.
Incorrect Answers
Answer B is incorrect because Using AWS Wavelength for edge-proximate inference integration would let the team place application components close to carrier edge locations to reduce network latency for mobile workloads. That is appropriate when a mobile GenAI experience requires very low latency near end users at the telecom edge. The scenario instead calls for jurisdiction or policy requires sensitive data processing to stay on premises while using AWS-consistent infrastructure.
Answer C is incorrect because store-and-forward integration during link outages has value because it can persist integration events locally and forward them after connectivity returns instead of dropping or blocking the business process. Its proper fit is when an edge or on-premises site must continue operating during temporary cloud disconnection. This question requires jurisdiction or policy requires sensitive data processing to stay on premises while using AWS-consistent infrastructure.
Answer D is incorrect because region-aware request routing for data residency can help a pipeline select a compliant regional processing path based on the data’s jurisdiction before FM access, especially when requests from different jurisdictions must remain within approved regions. The present requirement is jurisdiction or policy requires sensitive data processing to stay on premises while using AWS-consistent infrastructure, so it is not the best answer.
Question 12
A developer is hardening an FM-consumption workflow. An on-premises document system must send approved metadata to a VPC-hosted GenAI integration without traversing an open public endpoint. Which solution most directly resolves this issue? Choose ONE.
- compliance-aware model endpoint selection
- separate regional integration endpoints
- encrypted VPC-to-on-premises routing
- region-aware request routing for data residency
Correct Answer: C
Correct Answer
Answer C is correct because Select encrypted VPC-to-on-premises routing. Its purpose is to use private routed connectivity with encryption controls between cloud VPC resources and enterprise networks. The operational requirement is that cloud and on-premises components must exchange approved data over a controlled network path.
Incorrect Answers
Answer A is incorrect because compliance-aware model endpoint selection would be sound if the goal were data residency rules override convenience routing to any available model location; it can map each jurisdiction to an approved model and Region rather than automatically using cross-region inference. The requirement here is cloud and on-premises components must exchange approved data over a controlled network path.
Answer B is incorrect because separate regional integration endpoints can deploy independent endpoints and data stores per jurisdiction rather than sharing one global processing path, which matters when regional isolation is required for both data storage and FM request handling. The present case is driven by cloud and on-premises components must exchange approved data over a controlled network path.
Answer D is incorrect because region-aware request routing for data residency is useful when requests from different jurisdictions must remain within approved regions and can select a compliant regional processing path based on the data’s jurisdiction before FM access. The described workload requires cloud and on-premises components must exchange approved data over a controlled network path.
Question 13
A GenAI workload is moving from prototype to production. A factory can send maintenance narratives to the cloud only after employee identifiers and restricted asset IDs are removed on premises. Which design is the best fit for this requirement? Choose ONE.
- separate regional integration endpoints
- local redaction before cloud transfer
- AWS Wavelength for edge-proximate inference integration
- store-and-forward integration during link outages
Correct Answer: B
Correct Answer
Answer B is correct because The strongest answer is local redaction before cloud transfer. It works by helping to remove or tokenize restricted fields in the local environment before sending the remaining content to cloud GenAI services. The requirement is that raw regulated data cannot leave the local site but de-identified content may be processed in AWS.
Incorrect Answers
Answer A is incorrect because separate regional integration endpoints targets a neighboring concern. It can deploy independent endpoints and data stores per jurisdiction rather than sharing one global processing path when regional isolation is required for both data storage and FM request handling. The key requirement here is raw regulated data cannot leave the local site but de-identified content may be processed in AWS.
Answer C is incorrect because AWS Wavelength for edge-proximate inference integration can place application components close to carrier edge locations to reduce network latency for mobile workloads. Its best fit is where a mobile GenAI experience requires very low latency near end users at the telecom edge. The current case instead needs raw regulated data cannot leave the local site but de-identified content may be processed in AWS.
Answer D is incorrect because store-and-forward integration during link outages helps when an edge or on-premises site must continue operating during temporary cloud disconnection by allowing the pipeline to persist integration events locally and forward them after connectivity returns instead of dropping or blocking the business process. The described issue is raw regulated data cannot leave the local site but de-identified content may be processed in AWS.
Question 14
A regulated workflow requires deterministic preprocessing before inference. A customer master record has 60 fields, yet the cloud GenAI application is allowed to receive only 12 non-sensitive attributes. Which implementation best satisfies the requirement? Choose ONE.
- compliance-aware model endpoint selection
- approved-field synchronization across environments
- region-aware request routing for data residency
- store-and-forward integration during link outages
Correct Answer: B
Correct Answer
Answer B is correct because approved-field synchronization across environments is correct. It replicate only explicitly permitted attributes between on-premises and cloud systems. This matters because enterprise synchronization is required but sensitive columns must remain in the source environment. That directly addresses the scenario.
Incorrect Answers
Answer A is incorrect because compliance-aware model endpoint selection is useful when data residency rules override convenience routing to any available model location because it can map each jurisdiction to an approved model and Region rather than automatically using cross-region inference. The current requirement is enterprise synchronization is required but sensitive columns must remain in the source environment, which makes this a mismatch.
Answer C is incorrect because Using region-aware request routing for data residency would let the team select a compliant regional processing path based on the data’s jurisdiction before FM access. That is appropriate when requests from different jurisdictions must remain within approved regions. The scenario instead calls for enterprise synchronization is required but sensitive columns must remain in the source environment.
Answer D is incorrect because store-and-forward integration during link outages has value because it can persist integration events locally and forward them after connectivity returns instead of dropping or blocking the business process. Its proper fit is when an edge or on-premises site must continue operating during temporary cloud disconnection. This question requires enterprise synchronization is required but sensitive columns must remain in the source environment.
Question 15
During a production-readiness review, engineers identify a data-handling issue. A branch office can transfer archive documents to S3 only between midnight and 4 a.m. without affecting daytime business traffic. What should be used to meet this requirement? Choose ONE.
- AWS DataSync bandwidth-scheduled transfer
- AWS Wavelength for edge-proximate inference integration
- separate regional integration endpoints
- compliance-aware model endpoint selection
Correct Answer: A
Correct Answer
Answer A is correct because Choose AWS DataSync bandwidth-scheduled transfer. Its role is to schedule managed data movement during approved windows to control network impact between on-premises and AWS. The key requirement is that large on-premises datasets may move to AWS only during a limited nightly network window.
Incorrect Answers
Answer B is incorrect because AWS Wavelength for edge-proximate inference integration would be sound if the goal were a mobile GenAI experience requires very low latency near end users at the telecom edge; it can place application components close to carrier edge locations to reduce network latency for mobile workloads. The requirement here is large on-premises datasets may move to AWS only during a limited nightly network window.
Answer C is incorrect because separate regional integration endpoints can deploy independent endpoints and data stores per jurisdiction rather than sharing one global processing path, which matters when regional isolation is required for both data storage and FM request handling. The present case is driven by large on-premises datasets may move to AWS only during a limited nightly network window.
Answer D is incorrect because compliance-aware model endpoint selection belongs in workflows where data residency rules override convenience routing to any available model location because it can map each jurisdiction to an approved model and Region rather than automatically using cross-region inference. This question instead asks for large on-premises datasets may move to AWS only during a limited nightly network window.
Question 16
A cloud engineering team must correct a data-processing weakness. A gateway configuration and application update must move from development to staging to production with auditable approvals. What is the most appropriate implementation choice? Choose ONE.
- CodeDeploy blue-green release with rollback
- CloudWatch and X-Ray gateway observability
- CodeBuild automated integration tests
- CodePipeline promotion for GenAI integration releases
Correct Answer: D
Correct Answer
Answer D is correct because CodePipeline promotion for GenAI integration releases is appropriate because it will orchestrate source, build, test, approval, and deployment stages for repeatable enterprise releases. The requirement centers on a situation where GenAI integration changes need controlled promotion across environments.
Incorrect Answers
Answer A is incorrect because CodeDeploy blue-green release with rollback works for cases where a GenAI gateway update must minimize downtime and allow immediate reversal if health checks fail by helping to shift traffic between old and new application versions while retaining a fast rollback path. The present scenario needs GenAI integration changes need controlled promotion across environments.
Answer B is incorrect because CloudWatch and X-Ray gateway observability targets a neighboring concern. It can capture metrics, logs, and distributed traces across the enterprise integration boundary when operators need request-level visibility across gateway, function, and downstream GenAI calls. The key requirement here is GenAI integration changes need controlled promotion across environments.
Answer C is incorrect because CodeBuild automated integration tests is designed to run repeatable build-time tests against API contracts, prompt schemas, and integration behavior before deployment. That makes sense when a release should fail automatically when the GenAI component breaks an enterprise contract. The question is centered on GenAI integration changes need controlled promotion across environments.
Question 17
A solutions architect is reviewing how source data reaches an FM. A containerized integration service must not reach production if its build contains a critical vulnerable dependency. What should the developer implement? Choose ONE.
- AWS AppConfig externalized integration configuration
- CodeDeploy blue-green release with rollback
- CI security scan before deployment
- API Gateway centralized model abstraction
Correct Answer: C
Correct Answer
Answer C is correct because The requirement points to CI security scan before deployment. It is meant to add automated dependency, infrastructure, and policy checks to the pipeline before a GenAI component can be promoted. The scenario needs an approach where enterprise compliance requires releases with known critical security findings to be blocked.
Incorrect Answers
Answer A is incorrect because AWS AppConfig externalized integration configuration can store and deploy routing or feature configuration separately from application code with controlled rollout. It fits when model-routing rules change frequently and should not require a full code deployment. This scenario instead requires enterprise compliance requires releases with known critical security findings to be blocked, so it targets a different problem.
Answer B is incorrect because CodeDeploy blue-green release with rollback is useful when a GenAI gateway update must minimize downtime and allow immediate reversal if health checks fail because it can shift traffic between old and new application versions while retaining a fast rollback path. The current requirement is enterprise compliance requires releases with known critical security findings to be blocked, which makes this a mismatch.
Answer D is incorrect because Using API Gateway centralized model abstraction would let the team expose a stable enterprise API while backend model providers or versions can change behind the gateway. That is appropriate when client applications should not depend on provider-specific FM endpoint details. The scenario instead calls for enterprise compliance requires releases with known critical security findings to be blocked.
Question 18
A platform team is reviewing a production GenAI data pipeline. A release is technically healthy but produces responses that fail required contract and safety tests, so the pipeline must restore the last approved version. Which implementation provides the required behavior? Choose ONE.
- automated quality gate with rollback
- CodeBuild automated integration tests
- CloudWatch and X-Ray gateway observability
- API Gateway centralized model abstraction
Correct Answer: A
Correct Answer
Answer A is correct because automated quality gate with rollback matches the constraint. It can evaluate predefined GenAI integration tests after deployment and revert when acceptance thresholds fail. This is the intended approach when deployment success depends on behavioral quality as well as infrastructure health.
Incorrect Answers
Answer B is incorrect because CodeBuild automated integration tests is appropriate where a release should fail automatically when the GenAI component breaks an enterprise contract. It achieves that by helping to run repeatable build-time tests against API contracts, prompt schemas, and integration behavior before deployment. This scenario instead requires deployment success depends on behavioral quality as well as infrastructure health.
Answer C is incorrect because CloudWatch and X-Ray gateway observability would be sound if the goal were operators need request-level visibility across gateway, function, and downstream GenAI calls; it can capture metrics, logs, and distributed traces across the enterprise integration boundary. The requirement here is deployment success depends on behavioral quality as well as infrastructure health.
Answer D is incorrect because API Gateway centralized model abstraction can expose a stable enterprise API while backend model providers or versions can change behind the gateway, which matters when client applications should not depend on provider-specific FM endpoint details. The present case is driven by deployment success depends on behavioral quality as well as infrastructure health.
Question 19
An enterprise team is refining the input pipeline for a GenAI application. One internal application occasionally floods the GenAI gateway and causes throttling for every other team. Which change should the team make? Choose ONE.
- CodeDeploy blue-green release with rollback
- CloudWatch and X-Ray gateway observability
- API Gateway per-client throttling
- AWS AppConfig externalized integration configuration
Correct Answer: C
Correct Answer
Answer C is correct because Use API Gateway per-client throttling for this design. It will apply quotas or rate limits by client so one application cannot consume all shared GenAI capacity. That aligns with the need that many enterprise consumers share one gateway and noisy-neighbor behavior must be controlled.
Incorrect Answers
Answer A is incorrect because CodeDeploy blue-green release with rollback would improve another stage by enabling it to shift traffic between old and new application versions while retaining a fast rollback path. It fits when a GenAI gateway update must minimize downtime and allow immediate reversal if health checks fail. The stated problem is many enterprise consumers share one gateway and noisy-neighbor behavior must be controlled.
Answer B is incorrect because CloudWatch and X-Ray gateway observability works for cases where operators need request-level visibility across gateway, function, and downstream GenAI calls by helping to capture metrics, logs, and distributed traces across the enterprise integration boundary. The present scenario needs many enterprise consumers share one gateway and noisy-neighbor behavior must be controlled.
Answer D is incorrect because AWS AppConfig externalized integration configuration is designed to store and deploy routing or feature configuration separately from application code with controlled rollout. That makes sense when model-routing rules change frequently and should not require a full code deployment. The question is centered on many enterprise consumers share one gateway and noisy-neighbor behavior must be controlled.
Question 20
An application team is standardizing data before foundation-model inference. A shared GenAI API is changing a response field, and the platform team must detect incompatibility with current consumers before production. Which approach most directly addresses the stated constraint? Choose ONE.
- consumer-driven contract tests
- API Gateway centralized model abstraction
- CodeBuild automated integration tests
- AWS AppConfig externalized integration configuration
Correct Answer: A
Correct Answer
Answer A is correct because consumer-driven contract tests is the best fit because it can validate that a new gateway release continues to satisfy the request and response expectations of existing enterprise clients. The workflow needs a case where multiple dependent applications could break even when provider-side unit tests pass.
Incorrect Answers
Answer B is incorrect because API Gateway centralized model abstraction can expose a stable enterprise API while backend model providers or versions can change behind the gateway. It fits when client applications should not depend on provider-specific FM endpoint details. This scenario instead requires multiple dependent applications could break even when provider-side unit tests pass, so it targets a different problem.
Answer C is incorrect because CodeBuild automated integration tests is useful when a release should fail automatically when the GenAI component breaks an enterprise contract because it can run repeatable build-time tests against API contracts, prompt schemas, and integration behavior before deployment. The current requirement is multiple dependent applications could break even when provider-side unit tests pass, which makes this a mismatch.
Answer D is incorrect because AWS AppConfig externalized integration configuration addresses another concern by helping to store and deploy routing or feature configuration separately from application code with controlled rollout. It fits where model-routing rules change frequently and should not require a full code deployment. Here the deciding need is multiple dependent applications could break even when provider-side unit tests pass.
