AWS SAP-C02: AWS Certified Solutions Architect - Professional

Last checked October 22, 2026 against the official AWS exam page.

1. Question 1Design New Solutions

   A web application must survive the failure of an entire Availability Zone with minimal data loss for its relational database. Which configuration provides this?

   1. AA single-AZ instance with daily snapshots stored in one zone and restored manually after a failure
   2. BA read replica in the same AZ configured with asynchronous replication and promoted by hand during an outage
   3. CA Multi-AZ database deployment with synchronous replication to a standby in another AZ
   4. DStoring the database on a laptop under a desk connected to a consumer Wi-Fi network with a USB drive
   Show answer and explanationHide answer

   Correct answer: C

   Multi-AZ synchronous replication keeps a standby in another zone for automatic failover with minimal data loss. Single-AZ instances, same-AZ replicas and laptops do not survive zone failures.

2. Question 2Continuous Improvement for Existing Solutions

   A company's monthly AWS spend grows quickly. The team wants to see which accounts and services drive costs and be alerted to unusual spikes. Which combination helps most?

   1. ACost Explorer with cost allocation tags and Cost Anomaly Detection alerts
   2. BTurning off monitoring to reduce costs by disabling cost reports in every account
   3. CManually reviewing invoices once a year by a finance intern from PDF bills in shared folders
   4. DAsking each team to guess their spend from memory during quarterly planning meetings
   Show answer and explanationHide answer

   Correct answer: A

   Cost Explorer with tags and anomaly alerts identifies drivers and unusual spikes. Yearly invoice reviews, disabled monitoring and guessing do not provide timely insight.

3. Question 3Design Solutions for Organizational Complexity

   A company with several accounts wants centralized logging of all CloudTrail events in a separate account that application teams cannot modify. Which design meets this?

   1. ACloudTrail enabled separately in each account, with single-account trails writing to local S3 buckets that each account team manages
   2. BSharing the root user credentials of the management account with partner teams so they can read log files from the trail's bucket
   3. CAn organization trail delivering to an S3 bucket in a dedicated log archive account with restricted write access
   4. DDisabling CloudTrail in the application accounts during release windows to keep trail delivery costs and event volume low overall
   Show answer and explanationHide answer

   Correct answer: C

   An organization trail to a dedicated archive account centralizes logs and keeps them outside application-team control. Local buckets, disabled trails and shared root credentials fail the requirement.

4. Question 4Accelerate Workload Migration and Modernization

   A company is migrating a monolithic application to containers and wants to modernize incrementally by moving one function at a time behind an API gateway. Which pattern does this describe?

   1. AHard-coded routing in every client
   2. BStrangler fig pattern
   3. CBig-bang cutover
   4. DLift and shift without any changes
   Show answer and explanationHide answer

   Correct answer: B

   The strangler fig pattern replaces functionality incrementally behind a routing layer. Big-bang cutovers, unchanged lift-and-shift and hard-coded client routing do not describe incremental modernization.

5. Question 5Design New Solutions

   A serverless backend needs to process events from an SQS queue and write results to DynamoDB. Failed events must be retried a limited number of times and then captured for investigation. Which configuration fits?

   1. AA Lambda event source mapping with a dead-letter queue for failed batches and a bounded retry policy
   2. BDeleting failed events without record after their visibility timeout expires on the source queue, so no trace remains
   3. CA cron job that polls the queue every second with a sleep loop in a single process on a virtual machine, retrying each message
   4. DStoring failed events in a text file in /tmp on the function's disk, which is ephemeral and recycled between invocations
   Show answer and explanationHide answer

   Correct answer: A

   Lambda event source mapping with a DLQ and bounded retries captures repeated failures for analysis. Polling every second, local file storage and silent deletion do not meet the requirement.