SAA-C03 Design High-Performing Architectures Practice Test 4 — 15 Questions

Question 1

A DynamoDB-backed multi-tenant app experiences throttling. Most write traffic for tenant 'ACME' targets a single logical stream of events (you write items for ACME in near-real time). The table currently uses partition key = tenantId and sort key = eventTimestamp. CloudWatch shows partition-level throttling concentrated in the ACME partition. What design change most directly improves write throughput for the hottest tenant while still enabling efficient queries for recent events for that tenant?

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

Accepted Answer

Mitigate the hotspot by changing the partition key to include a shard value (for example, tenantId + '#' + shardId) and write using shardId. Query recent events by fanning out across ACME shards and merging results by eventTimestamp.. It directly addresses the partition-level throttling by introducing a shard key (e.g., tenantId + '#' + shardId) as the partition key, which distributes ACME's write load across multiple physical partitions. To query recent events for ACME, the application must fan out queries across all shards and merge results by eventTimestamp, which is efficient because each shard holds a subset of the data and the sort key remains eventTimestamp for ordering.

Answer

Add a Global Secondary Index (GSI) with the same partition key (tenantId) and eventTimestamp, and rely on the GSI to spread load.

Answer

Increase the table’s write capacity (or on-demand baseline) without changing the partition key, because DynamoDB will automatically balance hotspots.

Answer

Switch the sort key to a random value to prevent writes from landing on the same physical partition.

Question 2

A data processing application runs on a single EC2 instance and needs persistent block storage with sustained low-latency random read/write performance (high IOPS). Which storage choice is most appropriate?

Accepted Answer

EBS io2 provisioned IOPS SSD. Amazon EBS io2 Provisioned IOPS SSD volumes are designed for I/O-intensive workloads that require sustained, low-latency random read/write performance with high IOPS. They provide consistent performance by allowing you to provision a specific IOPS level (up to 256,000 IOPS per volume) and offer a 99.999% durability guarantee, making them ideal for a single EC2 instance needing persistent block storage.

Answer

Amazon S3 Standard

Answer

Amazon EFS for POSIX file sharing between multiple instances

Answer

EBS Throughput Optimized HDD (st1) storage

Question 3

A trading analytics system deploys multiple EC2 instances that exchange very frequent, low-latency, east-west messages. The application team wants the instances to be placed to minimize network latency and variability. Which AWS feature should they use?

Accepted Answer

EC2 Placement Groups with the "cluster" strategy. The cluster placement group is the correct choice because it places instances into a low-latency, high-bandwidth group within a single Availability Zone, which minimizes network latency and variability for east-west traffic. This strategy is specifically designed for applications that require very frequent, low-latency communication between instances, such as trading analytics systems.

Answer

EC2 Placement Groups with the "spread" strategy

Answer

Auto Scaling cooldown adjustments only

Answer

Switching the instances to a larger instance size without any placement group

Question 4

Your team serves static JavaScript and CSS files from an S3 origin through CloudFront. After a release, the CloudFront cache hit ratio dropped because clients keep re-downloading the same assets. What is the best next change to improve caching performance?

Accepted Answer

Update origin responses to include long-lived Cache-Control headers (for example, max-age) so CloudFront can cache objects. Setting long-lived Cache-Control headers (e.g., max-age=31536000) on static assets tells CloudFront to cache them at edge locations for an extended period. This reduces the number of requests forwarded to the S3 origin, improving the cache hit ratio and preventing clients from re-downloading unchanged assets on every visit.

Answer

Switch the S3 bucket to S3 Glacier so objects are not frequently accessed

Answer

Disable CloudFront compression to reduce CPU usage at the edge

Answer

Set CloudFront to forward all query strings to the origin to ensure the latest assets are returned

Question 5

A DynamoDB-backed multi-tenant app experiences throttling during a promotion. Most writes and reads target tenant "ACME" and use the same partition key value, causing a hot partition. Which design change most directly improves performance?

Accepted Answer

Add a "shard" component to the partition key (for example, tenantId + hashed bucket) to spread traffic across partitions. Adding a shard component to the partition key (e.g., appending a random or hash-based suffix to the tenant ID) distributes writes and reads for the same tenant across multiple physical partitions. This directly alleviates the hot partition caused by all ACME traffic hitting a single partition key value, allowing DynamoDB to utilize its full provisioned throughput across partitions.

Answer

Increase the table’s read capacity without changing the partition key

Answer

Switch all reads to strongly consistent reads to guarantee faster results

Answer

Store ACME data in S3 and query it directly to avoid DynamoDB throttling

Question 6

A service performs many repeated read requests for the same DynamoDB items. The reads are latency-sensitive, but the application can tolerate slightly stale data. Which AWS service is the best fit to reduce read latency?

Accepted Answer

Amazon DAX (DynamoDB Accelerator). Amazon DAX (DynamoDB Accelerator) is an in-memory cache specifically designed for DynamoDB. It reduces read latency from single-digit milliseconds to microseconds by caching frequently accessed items, and it supports eventually consistent reads, which aligns with the application's tolerance for slightly stale data. DAX handles repeated read requests without additional DynamoDB read capacity unit consumption, making it the optimal choice for this latency-sensitive workload.

Answer

Amazon S3 Select

Answer

Amazon SQS FIFO queue

Answer

AWS Lambda provisioned concurrency

Question 7

An application uses an Amazon Aurora cluster. The workload becomes read-heavy, but the team cannot change the database schema. They need higher read throughput while keeping writes on the primary. What should they do?

Accepted Answer

Create Aurora read replicas and use the reader endpoint for read traffic. Aurora read replicas are designed to offload read traffic from the primary instance, and the reader endpoint automatically load-balances connections across all replicas. Since the workload is read-heavy and the schema cannot change, adding read replicas directly increases read throughput without modifying the application's database schema. The reader endpoint ensures that read queries are directed to the replicas while writes continue to hit the primary instance.

Answer

Switch the cluster to a single-AZ Aurora configuration to reduce coordination overhead

Answer

Increase DynamoDB capacity units instead of modifying the database layer

Answer

Enable CloudFront caching for database queries to serve results from edge locations

Question 8

A latency-sensitive API is implemented with AWS Lambda. During traffic ramp-ups, users sometimes experience slow responses due to cold starts. The team wants to ensure fast initialization for a baseline level of concurrent requests. Which AWS feature should they use?

Accepted Answer

Lambda provisioned concurrency. Lambda Provisioned Concurrency keeps a specified number of execution environments initialized and ready to respond immediately, eliminating cold starts for those concurrent requests. This directly addresses the latency-sensitive API requirement during traffic ramp-ups by ensuring fast initialization for a baseline level of concurrency.

Answer

Increase reserved instances for EC2

Answer

Enable S3 event notifications for every request to the API

Answer

Decrease the function timeout to reduce execution variability

Question 9

A media company serves versioned JavaScript and CSS from an S3 origin through CloudFront. After a release, the cache hit ratio drops because the SPA sends an Authorization header and several tracking query strings on every request, even though the assets are public and identical for all users. Which changes would most improve cache efficiency without changing the content returned? Select three.

Accepted Answer

Create a CloudFront cache policy that excludes the Authorization header from the cache key when the assets do not require per-user authorization.. CloudFront's default behavior includes the Authorization header in the cache key, causing unique cache entries for each user even when the content is public. By creating a cache policy that excludes the Authorization header, CloudFront treats all requests for the same object as identical, dramatically improving the cache hit ratio without affecting the content served.

Answer

Place the S3 origin behind an Application Load Balancer so CloudFront can reuse more cached responses.

Answer

Enable S3 Transfer Acceleration to increase the cache hit ratio for repeated browser requests.

Question 10

An event-ingestion application writes telemetry to DynamoDB with partition key tenantId and sort key eventTime. During a promotion, one tenant generates 10 times the normal traffic. Dashboards repeatedly query the most recent items for that tenant, and they can tolerate slightly stale data. Which changes would most effectively reduce throttling and improve responsiveness? Select three.

Accepted Answer

Introduce a sharded partition key for the hot tenant and query the small shard set when reading recent data.. Introducing a sharded partition key (e.g., tenantId#shardId) for the hot tenant spreads its write traffic across multiple physical partitions, reducing throttling. When querying recent data, the dashboard can read from a small, fixed set of shards (e.g., 10 shards) and merge results, which is efficient and tolerates the slight staleness. This pattern directly addresses the hot partition issue without requiring application-level aggregation.

Answer

Increase only the sort-key cardinality while leaving the partition key unchanged.

Answer

Move the table to the Standard-IA table class because throttling is usually caused by storage class selection.

Question 11

A rendering service runs on a single EC2 instance and writes a large working set of metadata to disk using sustained random reads and writes. The data must persist across stops and restarts, and the team sees queue depth spikes when the job reaches peak throughput. Which changes should the team make? Select three.

Accepted Answer

Use an Amazon EBS io2 volume with provisioned IOPS for the metadata store.. An Amazon EBS io2 volume with provisioned IOPS is designed for I/O-intensive workloads with sustained random reads and writes, such as the rendering service's metadata store. The io2 volume type offers high durability (99.999%) and consistent low-latency performance, which directly addresses the queue depth spikes caused by peak throughput demands. Provisioned IOPS ensures the volume can handle the required random I/O without throttling, meeting the persistence requirement across stops and restarts.

Answer

Move the working set to Amazon EFS because it automatically stripes across Availability Zones.

Answer

Store the metadata in Amazon S3 because object storage is cheaper and supports random writes.

Question 12

A static website stores assets in S3 and is delivered through CloudFront. Analytics show low cache hit ratio, many origin fetches for the same JavaScript bundles, and elevated S3 GET request costs. Most requests include unnecessary cookies, and the text assets are uncompressed. Which changes should the team make? Select three.

Accepted Answer

Configure a CloudFront cache policy that excludes unnecessary cookies and headers from the cache key.. A CloudFront cache policy that excludes unnecessary cookies and headers from the cache key prevents CloudFront from creating multiple cache entries for the same object based on varying cookie values. This increases the cache hit ratio by ensuring that identical JavaScript bundles are served from the edge cache rather than triggering separate origin fetches for each unique request, thereby reducing S3 GET request costs.

Answer

Switch the origin to an Application Load Balancer so CloudFront can cache the assets more effectively.

Answer

Disable caching so viewers always retrieve the newest version directly from S3.

Question 13

A latency-sensitive telemetry service uses a custom TCP protocol on EC2 instances in private subnets. The service must preserve the client source IP for rate limiting, avoid HTTP header inspection, and keep per-request overhead as low as possible. Which changes should the team make? Select three.

Accepted Answer

Use a Network Load Balancer in front of the service.. A Network Load Balancer (NLB) operates at Layer 4 and preserves the client source IP by default when instances are registered as targets. This allows the telemetry service to use the original IP for rate limiting without requiring HTTP header inspection, which is critical for a custom TCP protocol. NLB also introduces minimal latency and low per-request overhead, making it ideal for latency-sensitive workloads.

Answer

Use an Application Load Balancer because path-based routing improves throughput for binary protocols.

Answer

Expose the service through API Gateway because it supports raw TCP and UDP pass-through.

Question 14

A customer portal uses Amazon Aurora MySQL. The application currently sends all SELECT queries to the writer instance endpoint. During traffic spikes, read latency increases, and the team wants the cluster to survive a writer failover without manual endpoint changes for the application. Which changes should the team make? Select three.

Accepted Answer

Replace the hard-coded DB instance endpoint with the Aurora reader endpoint for read traffic.. The Aurora reader endpoint automatically distributes read-only connections across all available Aurora Replicas, including the writer instance if it is configured to accept reads. This eliminates the need for manual endpoint changes during a failover, as the reader endpoint remains constant while the underlying instances change.

Answer

Keep sending all traffic to the writer endpoint because it always has the freshest data.

Answer

Replace Aurora with a single larger RDS instance and continue using the same read pattern.

Question 15

A low-latency market-data engine runs 10 EC2 instances that exchange small messages thousands of times per second. The team wants the lowest possible network latency and jitter, and they can tolerate single-AZ placement for this tier because another layer handles disaster recovery. Which changes should they make? Select three.

Accepted Answer

Use a cluster placement group for the instances.. A cluster placement group is designed for low-latency, high-throughput scenarios by ensuring instances are in close proximity within a single Availability Zone, which minimizes network latency and jitter. This directly meets the requirement for the lowest possible network latency and jitter for the market-data engine.

Answer

Use a spread placement group to maximize low-latency communication across the fleet.

Answer

Distribute the instances across multiple Availability Zones to reduce intra-cluster latency.