SAP-C02 Design for New Solutions Practice Test 8 — 15 Questions

Question 1

A company is designing a serverless data processing pipeline using AWS Lambda. The pipeline processes data from an Amazon Kinesis Data Stream. The Lambda function has a memory limit of 512 MB and a timeout of 5 minutes. The data volume is expected to increase significantly. Which TWO strategies should the company implement to improve throughput and reduce processing latency? (Choose TWO.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Increase the number of shards in the Kinesis data stream. Option A is correct because Kinesis Data Streams scales throughput by shard: each shard supports up to 1 MB/s or 1,000 records/s ingress, so adding shards raises the stream's ingest capacity and allows more Lambda invocations to run in parallel (one concurrent Lambda per shard per batch), directly improving throughput. Option B is correct because increasing the batch size in the Lambda event source mapping lets each invocation process more records per poll, reducing the number of Lambda invocations, per-invocation overhead, and downstream calls, which lowers processing latency and cost while keeping up with higher volume. Option C is not appropriate because switching to Amazon SQS changes the ingestion model and does not inherently increase throughput or reduce latency for a Kinesis-based pipeline. Option D is not correct because raising memory to 1024 MB only proportionally increases CPU and may speed a single invocation, but it does not address the shard-level parallelism bottleneck or batch efficiency. Option E is not correct because reserved concurrency caps the maximum concurrent invocations; for Kinesis event source mappings, concurrency is bounded by the number of shards, so raising reserved concurrency alone will not increase throughput beyond that limit.

Answer

Change the data source from Kinesis to an Amazon SQS queue

Answer

Increase the Lambda function memory to 1024 MB

Answer

Increase the Lambda function's reserved concurrency to a higher value

Question 2

A company is deploying a containerized application on Amazon EKS. The application requires persistent storage that can be shared across multiple pods in different Availability Zones. Which TWO storage solutions meet this requirement? (Choose TWO.)

Accepted Answer

Amazon FSx for Lustre. Amazon EFS (E) is correct because it is a fully managed, elastic NFS file system that supports ReadWriteMany access, allowing the same file system to be mounted concurrently by many pods across multiple Availability Zones in a region. Amazon FSx for Lustre (D) is also correct because it is a high-performance parallel file system that supports shared access from multiple compute nodes and can be linked to an S3 bucket, making it suitable for multi-pod, multi-AZ workloads. Amazon EBS with io2 volumes (A) is incorrect because EBS volumes are block storage attached to a single EC2 instance/AZ and cannot be shared across pods in different Availability Zones. Amazon S3 (B) is object storage accessed via APIs, not a POSIX file system that pods can mount as shared persistent storage. Amazon EC2 Instance Store (C) is ephemeral local storage tied to a single host and provides no cross-AZ sharing or persistence.

Answer

Amazon EBS with io2 volumes

Answer

Amazon S3

Answer

Amazon EC2 Instance Store

Question 3

A company is designing a serverless application using AWS Lambda. The function needs to process files uploaded to an S3 bucket and store metadata in DynamoDB. The solution must handle up to 1,000 concurrent invocations. Which configuration should be used to avoid throttling?

Accepted Answer

Set reserved concurrency to 1,000. Reserved concurrency guarantees that a specific amount of concurrent executions is set aside exclusively for a function, preventing it from being throttled by other functions in the same account. Setting it to 1,000 ensures the function can handle up to 1,000 concurrent invocations without competing for the account's shared concurrency pool. This directly addresses the requirement to avoid throttling under the specified load.

Answer

Request a concurrency limit increase from AWS Support

Answer

Enable provisioned concurrency

Answer

Use a dead-letter queue (DLQ) to retry throttled requests

Question 4

A company is designing a new multi-region disaster recovery solution for a critical database. The database runs on Amazon RDS for MySQL in us-east-1. The recovery point objective (RPO) is 1 second, and the recovery time objective (RTO) is 1 minute. Which strategy meets these requirements?

Accepted Answer

Multi-AZ RDS instance with a cross-Region read replica in us-west-2. A Multi-AZ RDS instance with a cross-Region read replica in us-west-2 can achieve an RPO of 1 second and an RTO of 1 minute. The cross-Region read replica uses asynchronous replication with a typical lag of less than 1 second, meeting the RPO. For RTO, you can promote the read replica to a standalone instance in under a minute, and the Multi-AZ configuration in the primary region ensures high availability during the promotion process.

Answer

Single-AZ RDS instance with cross-Region snapshot copy

Answer

Multi-AZ RDS instance in us-east-1

Answer

Amazon Aurora Global Database

Question 5

A company is designing a real-time analytics pipeline to process streaming data from IoT devices. The solution must be serverless and handle data transformation before storage. Which combination of services is most cost-effective?

Accepted Answer

Amazon Kinesis Data Streams, AWS Lambda, and Amazon Kinesis Data Firehose to Amazon S3. It combines Kinesis Data Streams for real-time ingestion, AWS Lambda for serverless data transformation, and Kinesis Data Firehose to buffer and deliver transformed data to Amazon S3. This architecture is serverless, scales automatically, and is cost-effective as you pay only for data throughput and compute time without provisioning any servers.

Answer

Amazon Kinesis Data Streams, Amazon EC2 instances for transformation, and Amazon S3

Answer

Amazon Simple Queue Service (SQS), AWS Lambda, and Amazon S3

Answer

Amazon Kinesis Data Analytics for real-time SQL and Amazon S3

Question 6

A company is designing a new application that must meet PCI DSS compliance requirements. The application will process credit card transactions and store encrypted data. Which AWS service should be used to manage the encryption keys?

Accepted Answer

AWS CloudHSM. AWS CloudHSM is the correct choice because PCI DSS requires that encryption keys used to protect cardholder data be stored in a hardware security module (HSM) that is FIPS 140-2 Level 3 validated, and that the customer retains sole control over the keys. CloudHSM provides dedicated, single-tenant HSM appliances that you manage directly, ensuring you meet the key management and physical security requirements of PCI DSS without sharing the HSM with other AWS customers.

Answer

AWS Key Management Service (KMS)

Answer

AWS Secrets Manager

Answer

Amazon S3 server-side encryption (SSE-S3)

Question 7

A company is designing a new hybrid cloud architecture that extends on-premises storage to AWS. The solution must provide low-latency access to frequently accessed data and use AWS storage for backup. Which TWO services should be used together?

Accepted Answer

AWS Storage Gateway (File Gateway). AWS Storage Gateway (File Gateway) [B] is correct because it presents an on-premises NFS/SMB file interface backed by Amazon S3, caching frequently accessed data locally for low-latency access while transparently storing the primary data in AWS. Amazon S3 [E] is correct because it serves as the durable, scalable cloud storage backend for the File Gateway and provides the target for backup data. Together they satisfy the hybrid requirement of local low-latency access plus AWS-based backup. Amazon EFS [A] is a cloud-only managed NFS file system and does not extend on-premises storage or provide local caching. AWS Snowball Edge [C] is a physical data-transfer device for bulk migrations, not a low-latency hybrid storage solution. AWS Direct Connect [D] is a dedicated network connection that could complement the design but is not a storage service and does not itself provide the required storage or backup capability.

Answer

Amazon EFS

Answer

AWS Snowball Edge

Answer

AWS Direct Connect

Question 8

A company is designing a new serverless application using AWS Lambda. The function needs to access an Amazon RDS database. Which THREE practices should be followed to avoid connection exhaustion?

Accepted Answer

Use Amazon RDS Proxy to pool connections. Option C is correct because Amazon RDS Proxy sits between Lambda and RDS, maintaining a managed pool of database connections so that many concurrent Lambda invocations share a small number of actual database connections, directly preventing connection exhaustion. Option D is correct because opening a connection only when needed and closing it after each invocation avoids leaving idle connections open, which is essential when Lambda scales to many concurrent execution environments. Option E is correct because raising the max_connections value in the RDS parameter group (subject to instance memory limits) increases the ceiling of available connections, giving headroom when many Lambda instances connect simultaneously. Option A is not appropriate here because a global variable only reuses a connection within a single warm execution environment; it does not solve exhaustion across the many concurrent environments Lambda creates. Option B is incorrect because assigning a static IP address to Lambda does not reduce the number of database connections and is not a connection-management practice.

Answer

Store the database connection in a global variable to reuse across invocations

Answer

Assign a static IP address to the Lambda function

Question 9

A company is designing a new data processing pipeline that must transform data from JSON to Parquet format. The pipeline should run daily and handle data up to 10 GB. Which TWO AWS services can be used to perform this transformation?

Accepted Answer

AWS Glue. AWS Glue (C) is correct because it is a fully managed, serverless ETL service whose built-in transforms and DynamicFrame/Spark engine can read JSON and write Parquet, and its crawlers and jobs are well suited to a scheduled daily batch pipeline handling up to 10 GB. Amazon Athena (D) is correct because CTAS (CREATE TABLE AS SELECT) queries can read JSON data from Amazon S3 and write the results back to S3 in Parquet format, and Athena can be scheduled via Amazon EventBridge or orchestrated with Step Functions for a daily run. Amazon EMR (A) could technically run Spark/Hive to convert JSON to Parquet, but it is not one of the marked answers and requires provisioning and managing a cluster, which is heavier than the managed options intended here. AWS Lambda (B) is not appropriate because its 15-minute maximum execution time and 10 GB memory limit make it unsuitable for reliably transforming 10 GB datasets in a batch job. Amazon Kinesis Data Firehose (E) is a streaming delivery service that can convert JSON to Parquet via record format conversion, but it is designed for continuous streaming ingestion rather than a daily batch transformation pipeline.

Answer

Amazon EMR

Answer

AWS Lambda

Answer

Amazon Kinesis Data Firehose

Question 10

A company is designing a new microservices platform on AWS. The platform consists of 50 microservices, each running in its own Amazon ECS service on AWS Fargate. The services communicate via REST APIs. The company wants to implement a service mesh to handle traffic routing, observability, and security (mTLS). They also need to meet compliance requirements that all traffic between services must be encrypted and logged. The solution must be fully managed and reduce operational overhead. After implementing the service mesh, the operations team notices that latency between services has increased by 20%, and some services are experiencing connection timeouts. The team has enabled mTLS and distributed tracing. Which course of action should the team take to diagnose and resolve the latency issues?

Accepted Answer

Check the Envoy proxy resource limits in the App Mesh configuration and increase the CPU and memory allocated to the sidecar proxies.. The latency increase and connection timeouts are likely due to resource constraints on the Envoy sidecar proxies. When mTLS and distributed tracing are enabled, the proxies consume more CPU and memory. If the ECS task definition does not allocate sufficient resources to the sidecar, the proxy may become a bottleneck, causing latency and timeouts. Increasing the CPU and memory allocated to the sidecar proxies can resolve this.

Answer

Use AWS Cloud Map for service discovery instead of App Mesh.

Answer

Replace the service mesh with VPC peering and security groups, and use direct HTTP calls.

Answer

Convert the microservices to AWS Lambda functions and use API Gateway.

Question 11

A company is designing a multi-region active-active application using Amazon Aurora Global Database. The application writes to a custom domain endpoint that routes to the primary cluster. To minimize write latency, the application should write to the nearest region. Which configuration should the solutions architect use?

Accepted Answer

Use Amazon DynamoDB global tables instead of Aurora Global Database, as DynamoDB supports multi-region writes.. Amazon Aurora Global Database does not support multi-region writes; it has a single primary (writer) cluster and multiple read-only secondary regions. Amazon DynamoDB global tables, however, natively support active-active multi-region writes, allowing the application to write to the nearest region and achieve low write latency. Therefore, DynamoDB global tables are the correct service for this requirement.

Answer

Configure Aurora Global Database with multiple primary clusters, each in a different region, and use Route 53 to route writes.

Answer

Use Aurora cross-region read replicas and failover to a secondary region for writes.

Answer

Use Route 53 latency-based routing to direct writes to the nearest region. Each region has its own Aurora cluster.

Question 12

A data analytics company is building a real-time streaming pipeline using Amazon Kinesis Data Streams. The data is consumed by multiple consumer applications, each with different processing requirements. The company wants to ensure that each consumer can process records independently without affecting others and can reprocess data from a specific point in time. Which feature should the company use?

Accepted Answer

Use Enhanced Fan-Out with a timestamp to start reading.. Enhanced Fan-Out (EFO) provides each consumer with a dedicated 2 MB/second read throughput per shard, ensuring independent processing without contention. By using the SubscribeToShard API with a starting position specified via a timestamp, consumers can reprocess data from a specific point in time, meeting the requirement exactly.

Answer

Increase the data retention period to 365 days.

Answer

Use resharding to increase the number of shards.

Answer

Use the Kinesis Client Library (KCL) with checkpointing.

Question 13

A company is deploying a containerized application on Amazon ECS with Fargate. The application needs to store session state data that must be highly available and low latency. The data is accessed frequently and can be recreated if lost. Which storage solution should the solutions architect recommend?

Accepted Answer

Store session state in Amazon ElastiCache for Redis.. Amazon ElastiCache for Redis is the ideal choice for storing session state data in a containerized ECS with Fargate environment because it provides in-memory caching with sub-millisecond latency, high availability through replication and automatic failover, and supports data persistence. Since the session data can be recreated if lost, the ephemeral nature of Redis is acceptable, and its low-latency access pattern perfectly matches the frequent read/write requirements of session state.

Answer

Store session state in Amazon DynamoDB.

Answer

Store session state in Amazon S3.

Answer

Store session state in Amazon EFS.

Question 14

A company is designing a serverless application using AWS Lambda functions that process messages from an Amazon SQS queue. The Lambda function sometimes experiences throttling, causing messages to be sent to the dead-letter queue (DLQ). The company wants to minimize throttling and ensure that messages are processed in order. What should the solutions architect do?

Accepted Answer

Use a FIFO SQS queue and configure the Lambda function with reserved concurrency.. Using an SQS FIFO queue guarantees strict message ordering, and reserved concurrency on the Lambda function prevents throttling by allocating a fixed number of concurrent executions exclusively for this function. This combination ensures messages are processed in order without being throttled, avoiding unnecessary DLQ deliveries.

Answer

Use a standard SQS queue and configure a Lambda function with a higher concurrency limit.

Answer

Use an SQS FIFO queue with provisioned concurrency on the Lambda function.

Answer

Increase the batch size in the Lambda event source mapping and use a standard queue.

Question 15

A company is hosting a static website on Amazon S3. The website uses JavaScript to make API calls to a backend API hosted on Amazon API Gateway. The company wants to reduce latency for users worldwide. Which combination of AWS services should the solutions architect use? (Choose two.)

Accepted Answer

Use CloudFront with API Gateway as an origin.. Option A is correct because configuring Amazon CloudFront with API Gateway as an origin caches and serves API responses from edge locations, reducing latency for the JavaScript API calls made by users worldwide. Option D is correct because using CloudFront for both the S3 static website and the API Gateway endpoint delivers all content (HTML/JS and API responses) from AWS edge locations, minimizing round-trip latency globally. Option B is not appropriate because AWS Global Accelerator optimizes TCP/UDP traffic to regional endpoints and does not cache or accelerate S3 static website content the way CloudFront does. Option C is unnecessary because Lambda@Edge is used to run custom logic at edge locations, not to reduce latency for standard API Gateway calls, which CloudFront already accelerates. Option E does not belong because S3 Transfer Acceleration speeds up uploads to a bucket, not downloads of a static website or API responses to global users.

Answer

Use AWS Global Accelerator for the S3 website.

Answer

Use Lambda@Edge to process API calls at the edge.

Answer

Enable S3 Transfer Acceleration on the bucket.