Amazon AWS Certified Solutions Architect - Associate SAA-C03 Bundle

The option Configure Route 53 failover records and suitable health checks addresses a separate condition: traffic must leave a failed primary region and reach a healthy secondary region. Health-aware DNS can change regional answer selection. The actual requirement is a client requires static anycast public entry and fast health-based regional steering. Recommended corrective action: Use AWS Global Accelerator endpoints and health evaluation. Accelerator routes clients through static addresses to healthy regions.

Account for DNS TTL and resolver caching in the failover design would help if dNS clients keep using a failed region after the failover record changes. Client cache behavior controls DNS failover observation. That is not the symptom described here: a client requires static anycast public entry and fast health-based regional steering. The matching response is Use AWS Global Accelerator endpoints and health evaluation, because accelerator routes clients through static addresses to healthy regions.

For the alternative failure requests should go to the closest healthy endpoint by network latency, the control Configure Route 53 latency routing with endpoint health checks is relevant: latency policies choose regional targets using measured latency. Here the evidence instead is a client requires static anycast public entry and fast health-based regional steering. The selected technical mechanism is Use AWS Global Accelerator endpoints and health evaluation: accelerator routes clients through static addresses to healthy regions.

For the alternative failure developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions, the control Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required is relevant: default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints. Here the evidence instead is one DynamoDB table receives writes from a single region and the app requires regional fault isolation. The selected technical mechanism is Use DynamoDB's managed regional availability and design client retries: dynamoDB service architecture spans multiple Availability Zones within a Region.

Enable DynamoDB PITR and test restoration to a new table solves a different operational problem where the application requires point-in-time recovery from accidental record deletion. PITR provides historical recovery within its retention window. It does not directly address one DynamoDB table receives writes from a single region and the app requires regional fault isolation. The case calls for Use DynamoDB's managed regional availability and design client retries, since dynamoDB service architecture spans multiple Availability Zones within a Region.

The competing option Implement exponential backoff with jitter and capacity adjustments would be useful when an application has unexpected retry storms under provisioned throughput throttling. Retry discipline mitigates temporary throughput contention. The present issue is one DynamoDB table receives writes from a single region and the app requires regional fault isolation; it is more directly addressed by Use DynamoDB's managed regional availability and design client retries. DynamoDB service architecture spans multiple Availability Zones within a Region.

For the alternative failure developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions, the control Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required is relevant: default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints. Here the evidence instead is a global customer app needs active reads and writes in two regions. The selected technical mechanism is Configure DynamoDB global tables with application conflict considerations: global tables replicate data across regions with multi-region writes.

Enable DynamoDB PITR and test restoration to a new table solves a different operational problem where the application requires point-in-time recovery from accidental record deletion. PITR provides historical recovery within its retention window. It does not directly address a global customer app needs active reads and writes in two regions. The case calls for Configure DynamoDB global tables with application conflict considerations, since global tables replicate data across regions with multi-region writes.

The option Use DynamoDB's managed regional availability and design client retries addresses a separate condition: one DynamoDB table receives writes from a single region and the app requires regional fault isolation. DynamoDB service architecture spans multiple Availability Zones within a Region. The actual requirement is a global customer app needs active reads and writes in two regions. Recommended corrective action: Configure DynamoDB global tables with application conflict considerations. Global tables replicate data across regions with multi-region writes.

Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required solves a different operational problem where developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions. Default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints. It does not directly address the application requires point-in-time recovery from accidental record deletion. The case calls for Enable DynamoDB PITR and test restoration to a new table, since pITR provides historical recovery within its retention window.

The option Configure DynamoDB global tables with application conflict considerations addresses a separate condition: a global customer app needs active reads and writes in two regions. Global tables replicate data across regions with multi-region writes. The actual requirement is the application requires point-in-time recovery from accidental record deletion. Recommended corrective action: Enable DynamoDB PITR and test restoration to a new table. PITR provides historical recovery within its retention window.

Implement exponential backoff with jitter and capacity adjustments would help if an application has unexpected retry storms under provisioned throughput throttling. Retry discipline mitigates temporary throughput contention. That is not the symptom described here: the application requires point-in-time recovery from accidental record deletion. The matching response is Enable DynamoDB PITR and test restoration to a new table, because pITR provides historical recovery within its retention window.

The competing option Use DynamoDB's managed regional availability and design client retries would be useful when one DynamoDB table receives writes from a single region and the app requires regional fault isolation. DynamoDB service architecture spans multiple Availability Zones within a Region. The present issue is an application has unexpected retry storms under provisioned throughput throttling; it is more directly addressed by Implement exponential backoff with jitter and capacity adjustments. Retry discipline mitigates temporary throughput contention.

The option Enable DynamoDB PITR and test restoration to a new table addresses a separate condition: the application requires point-in-time recovery from accidental record deletion. PITR provides historical recovery within its retention window. The actual requirement is an application has unexpected retry storms under provisioned throughput throttling. Recommended corrective action: Implement exponential backoff with jitter and capacity adjustments. Retry discipline mitigates temporary throughput contention.

For the alternative failure developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions, the control Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required is relevant: default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints. Here the evidence instead is an application has unexpected retry storms under provisioned throughput throttling. The selected technical mechanism is Implement exponential backoff with jitter and capacity adjustments: retry discipline mitigates temporary throughput contention.

The option Implement exponential backoff with jitter and capacity adjustments addresses a separate condition: an application has unexpected retry storms under provisioned throughput throttling. Retry discipline mitigates temporary throughput contention. The actual requirement is developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions. Recommended corrective action: Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required. Default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints.

Enable DynamoDB PITR and test restoration to a new table would help if the application requires point-in-time recovery from accidental record deletion. PITR provides historical recovery within its retention window. That is not the symptom described here: developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions. The matching response is Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required, because default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints.

For the alternative failure a global customer app needs active reads and writes in two regions, the control Configure DynamoDB global tables with application conflict considerations is relevant: global tables replicate data across regions with multi-region writes. Here the evidence instead is developers assume DynamoDB global tables with default MREC provide strongly consistent reads across all AWS Regions. The selected technical mechanism is Design for MREC asynchronous replication lag or deliberately select supported MRSC at table creation when global strong reads are required: default MREC replication is eventual; MRSC is a distinct global-table mode with separate configuration constraints.

Configure a DLQ with inspected redrive policy would help if a message repeatedly fails and blocks successful work. Poison-message isolation prevents endless retries. That is not the symptom described here: a Lambda worker is unavailable for minutes while jobs arrive. The matching response is Store jobs in SQS and retry consumers when service resumes, because a durable queue buffers failures independently from producers.

Configure a supported Lambda on-failure destination or DLQ solves a different operational problem where lambda receives asynchronous invocation errors but no failure destination is configured. Failure destinations preserve records for later handling. It does not directly address a Lambda worker is unavailable for minutes while jobs arrive. The case calls for Store jobs in SQS and retry consumers when service resumes, since a durable queue buffers failures independently from producers.

The competing option Tune batching, concurrency and shard capacity within service constraints would be useful when an event source mapping cannot keep up with a high-volume stream. Stream consumption throughput depends on worker and shard configuration. The present issue is a Lambda worker is unavailable for minutes while jobs arrive; it is more directly addressed by Store jobs in SQS and retry consumers when service resumes. A durable queue buffers failures independently from producers.

Store jobs in SQS and retry consumers when service resumes solves a different operational problem where a Lambda worker is unavailable for minutes while jobs arrive. A durable queue buffers failures independently from producers. It does not directly address a message repeatedly fails and blocks successful work. The case calls for Configure a DLQ with inspected redrive policy, since poison-message isolation prevents endless retries.

The competing option Configure a supported Lambda on-failure destination or DLQ would be useful when lambda receives asynchronous invocation errors but no failure destination is configured. Failure destinations preserve records for later handling. The present issue is a message repeatedly fails and blocks successful work; it is more directly addressed by Configure a DLQ with inspected redrive policy. Poison-message isolation prevents endless retries.

The option Tune batching, concurrency and shard capacity within service constraints addresses a separate condition: an event source mapping cannot keep up with a high-volume stream. Stream consumption throughput depends on worker and shard configuration. The actual requirement is a message repeatedly fails and blocks successful work. Recommended corrective action: Configure a DLQ with inspected redrive policy. Poison-message isolation prevents endless retries.

The competing option Tune batching, concurrency and shard capacity within service constraints would be useful when an event source mapping cannot keep up with a high-volume stream. Stream consumption throughput depends on worker and shard configuration. The present issue is workers process a queue message and crash before deletion; it is more directly addressed by Make processing idempotent because message becomes visible again. At-least-once retries can repeat side effects.

The option Store jobs in SQS and retry consumers when service resumes addresses a separate condition: a Lambda worker is unavailable for minutes while jobs arrive. A durable queue buffers failures independently from producers. The actual requirement is workers process a queue message and crash before deletion. Recommended corrective action: Make processing idempotent because message becomes visible again. At-least-once retries can repeat side effects.

Configure a DLQ with inspected redrive policy would help if a message repeatedly fails and blocks successful work. Poison-message isolation prevents endless retries. That is not the symptom described here: workers process a queue message and crash before deletion. The matching response is Make processing idempotent because message becomes visible again, because at-least-once retries can repeat side effects.

The competing option Configure a DLQ with inspected redrive policy would be useful when a message repeatedly fails and blocks successful work. Poison-message isolation prevents endless retries. The present issue is lambda receives asynchronous invocation errors but no failure destination is configured; it is more directly addressed by Configure a supported Lambda on-failure destination or DLQ. Failure destinations preserve records for later handling.

Tune batching, concurrency and shard capacity within service constraints would help if an event source mapping cannot keep up with a high-volume stream. Stream consumption throughput depends on worker and shard configuration. That is not the symptom described here: lambda receives asynchronous invocation errors but no failure destination is configured. The matching response is Configure a supported Lambda on-failure destination or DLQ, because failure destinations preserve records for later handling.

For the alternative failure a Lambda worker is unavailable for minutes while jobs arrive, the control Store jobs in SQS and retry consumers when service resumes is relevant: a durable queue buffers failures independently from producers. Here the evidence instead is lambda receives asynchronous invocation errors but no failure destination is configured. The selected technical mechanism is Configure a supported Lambda on-failure destination or DLQ: failure destinations preserve records for later handling.

Configure a DLQ with inspected redrive policy would help if a message repeatedly fails and blocks successful work. Poison-message isolation prevents endless retries. That is not the symptom described here: an event source mapping cannot keep up with a high-volume stream. The matching response is Tune batching, concurrency and shard capacity within service constraints, because stream consumption throughput depends on worker and shard configuration.

For the alternative failure lambda receives asynchronous invocation errors but no failure destination is configured, the control Configure a supported Lambda on-failure destination or DLQ is relevant: failure destinations preserve records for later handling. Here the evidence instead is an event source mapping cannot keep up with a high-volume stream. The selected technical mechanism is Tune batching, concurrency and shard capacity within service constraints: stream consumption throughput depends on worker and shard configuration.

Store jobs in SQS and retry consumers when service resumes solves a different operational problem where a Lambda worker is unavailable for minutes while jobs arrive. A durable queue buffers failures independently from producers. It does not directly address an event source mapping cannot keep up with a high-volume stream. The case calls for Tune batching, concurrency and shard capacity within service constraints, since stream consumption throughput depends on worker and shard configuration.