I85PREP
⌘ K
AWS Solutions Architect Professional study materials
AWS Solutions Architect Professional study resources

AWS Solutions Architect Professional practice questions

One question at a time. Build your AWS Solutions Architect Professional understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company runs a production MySQL database on Amazon RDS and wants automatic failover to a standby instance in a different Availability Zone when the primary fails. Which RDS deployment option provides this?

  1. An Amazon Aurora global database with two Regions
  2. A daily automated snapshot copied to another Region
  3. A read replica in a second Availability Zone
  4. A Multi-AZ DB instance deployment Correct answer.

Explanation

Multi-AZ deployments keep a synchronously replicated standby in a different Availability Zone and automatically repoint the DNS endpoint during failover. Read replicas use asynchronous replication and must be promoted manually, and snapshots restore to a new instance rather than failing over automatically.

Question 2Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company wants 10% of user traffic to reach a new version of its web application while 90% continues to the current version, with both versions running at the same time. Which Route 53 routing policy should be used?

  1. Failover routing
  2. Geolocation routing
  3. Multivalue answer routing
  4. Weighted routing Correct answer.

Explanation

Weighted routing associates a numeric weight with each record so that DNS responses are distributed in the specified proportions, which is the standard way to split traffic between two live versions. Failover routing sends all traffic to one record unless a health check fails, and geolocation and multivalue policies split traffic by user location or at random rather than by a fixed percentage.

Question 3Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

When an Amazon SQS consumer receives a message, the message remains in the queue but is hidden from other consumers for a period. Which setting controls the length of that period?

  1. The message retention period
  2. The delay queue setting
  3. The receive request attempt ID
  4. The visibility timeout Correct answer.

Explanation

The visibility timeout is the window during which a received message is hidden from other consumers; if it is not deleted within that window it becomes visible again for redelivery. Retention period governs how long an undeleted message stays in the queue, and delay queues postpone the first delivery of new messages.

Question 4Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An Application Load Balancer target group has health checks enabled. What happens to a target that repeatedly fails its health check?

  1. The load balancer terminates the target instance
  2. The target is placed in a security group quarantine
  3. The Auto Scaling group is scaled in by one instance
  4. The load balancer stops routing new requests to that target Correct answer.

Explanation

Targets that fail health checks are marked unhealthy and removed from rotation, so the load balancer sends new requests only to healthy targets until the target passes again. The load balancer itself only reports health; replacing or terminating instances is the responsibility of the Auto Scaling group.

Question 5Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

Which prerequisite must be satisfied before enabling S3 Cross-Region Replication from a source bucket to a destination bucket?

  1. Versioning must be enabled on both buckets Correct answer.
  2. Both buckets must use the same storage class
  3. Transfer Acceleration must be enabled on the source bucket
  4. The buckets must reside in the same AWS account

Explanation

Replication operates on object versions, so versioning must be enabled on the source bucket and on the destination bucket. The buckets may be in different accounts, may use different storage classes, and do not require Transfer Acceleration.

Question 6Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An application's load varies unpredictably, and the team wants Auto Scaling group capacity to follow a metric such as average CPU utilization around a chosen value. Which scaling policy type should they use?

  1. Manual capacity adjustment
  2. Simple scaling with a cooldown
  3. Scheduled scaling
  4. Target tracking scaling Correct answer.

Explanation

Target tracking scaling creates and manages the CloudWatch alarms needed to keep a metric at or near a target value, which suits fluctuating demand with no predictable schedule. Scheduled scaling is bound to known time windows, and simple scaling reacts to a single threshold with cooldowns rather than continuously adjusting capacity.

Question 7Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

Which AWS service provides static anycast IP addresses that carry user traffic over the AWS global network to the nearest healthy endpoint in one or more Regions?

  1. An Application Load Balancer with cross-zone balancing
  2. AWS Global Accelerator Correct answer.
  3. Amazon CloudFront with origin groups
  4. Amazon Route 53 latency-based routing

Explanation

Global Accelerator advertises two static anycast IP addresses that onboard traffic at the nearest AWS edge and travel the AWS backbone to the optimal healthy endpoint. CloudFront accelerates and caches HTTP content but relies on DNS resolution and does not provide static anycast entry IPs.

Question 8Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A read-heavy reporting workload is overwhelming a single RDS for PostgreSQL instance. Which change offloads read traffic with the least application redesign?

  1. Enable Performance Insights with a longer retention period on the instance
  2. Convert the DB instance to a Multi-AZ deployment with a synchronous standby
  3. Add read replicas and route reads to them Correct answer.
  4. Increase the automated backup retention period to the maximum allowed

Explanation

Read replicas receive asynchronously replicated copies of the data and can serve read-only queries, which shifts reporting load off the primary with only a connection-endpoint change. Multi-AZ standbys do not accept read traffic, and monitoring or backup settings do not add read capacity.

Question 9Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company needs a relational database that can fail over to a second Region with a recovery point objective of about one second and a recovery time objective under one minute. Which design meets both targets?

  1. RDS for MySQL with automated backups replicated to the second Region
  2. An Amazon Aurora global database with a secondary Region Correct answer.
  3. RDS for MySQL Multi-AZ with cross-Region snapshot copies every hour
  4. Aurora MySQL with a cross-Region read replica created from a snapshot

Explanation

Aurora global databases replicate storage continuously to secondary Regions with typical replication lag around one second and support promotion of a secondary Region in under a minute. Snapshot-based or backup-based copies are periodic, so their RPO equals the copy interval and their RTO includes a restore operation.

Question 10Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An order service publishes to an Amazon SQS FIFO queue. Messages for the same customer must be processed in order, while different customers should be processed in parallel. What should the producer set on each message?

  1. A zero-second delivery delay on every message
  2. One dedicated FIFO queue per customer
  3. A distinct MessageDeduplicationId per customer
  4. A distinct MessageGroupId per customer Correct answer.

Explanation

A FIFO queue delivers messages that share a message group ID in strict order, one in-flight message per group, while different groups are consumed concurrently. The deduplication ID only suppresses duplicate sends within the deduplication interval and does not control ordering.

Question 11Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

During scale-in, an application must finish in-flight work and deregister itself from a service registry before an EC2 instance is terminated. Which Auto Scaling feature supports this?

  1. Instance refresh with a minimum healthy percentage
  2. A scale-in lifecycle hook with a heartbeat timeout Correct answer.
  3. A scheduled scaling action
  4. Termination protection on the launch template

Explanation

A scale-in lifecycle hook pauses the instance in the Terminating:Wait state so custom actions can run, and the heartbeat timeout bounds how long the instance may remain there. Instance refresh replaces instances in a rolling fashion but does not pause individual terminations for custom scripts.

Question 12Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

Users in North America and Europe need low-latency reads and writes to the same DynamoDB table, and each Region must keep accepting writes if the other becomes unavailable. Which solution should be used?

  1. DynamoDB global tables replicating between the two Regions Correct answer.
  2. A DynamoDB table with on-demand capacity and a read replica in Europe
  3. DynamoDB Accelerator (DAX) clusters in both Regions
  4. A nightly export of the table to Amazon S3 in each Region

Explanation

Global tables provide multi-Region, multi-active replication so each Region serves local reads and writes and continues operating independently during a Regional outage. DAX caches reads within a Region but does not replicate data across Regions or accept writes in a second Region.

Question 13Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A team deploys a Network Load Balancer with targets registered in three Availability Zones. By default, how is cross-zone load balancing configured on this load balancer type?

  1. Disabled and not configurable
  2. Enabled by default
  3. Enabled only for TLS listeners
  4. Disabled by default Correct answer.

Explanation

On Network Load Balancers, cross-zone load balancing is off by default and each node routes traffic only to targets in its own Availability Zone; it can be enabled per target group. Application Load Balancers, by contrast, have it enabled by default.

Question 14Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company runs a primary site in us-east-1 and a standby site in eu-west-1 using Route 53 failover routing. Traffic does not shift when the primary becomes unavailable. What is the most likely cause?

  1. The standby record uses a weighted routing policy
  2. The hosted zone is a private hosted zone
  3. No health check is associated with the primary record Correct answer.
  4. The record TTL is set too low

Explanation

Failover routing only diverts traffic when the primary record is found unhealthy, which requires an associated Route 53 health check. Without one, Route 53 continues to answer with the primary record regardless of the site's actual state.

Question 15Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A Lambda function serves both a latency-sensitive API and a nightly batch job. During batch runs the API is throttled because the function reaches the account concurrency limit. What is the most direct fix?

  1. Move the function into a VPC
  2. Configure reserved concurrency on the API function Correct answer.
  3. Increase the function's memory size
  4. Raise the function's timeout value

Explanation

Reserved concurrency carves out a guaranteed share of the account's concurrency pool for a function, so the batch workload cannot consume capacity needed by the API, and it also caps that function's own maximum concurrency. Memory, VPC placement, and timeout do not partition the concurrency pool between functions.

Question 16Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A workload runs on Spot Instances and can tolerate interruptions, but the team wants to minimize how often interruptions occur across instance families and Availability Zones. Which Spot allocation strategy should be used?

  1. lowestPrice
  2. diversified
  3. instancePoolsToUseCount set to 1
  4. capacity-optimized Correct answer.

Explanation

The capacity-optimized strategy launches instances from the pools with the deepest available capacity, which correlates with a lower frequency of Spot interruptions. lowestPrice concentrates instances in the cheapest pools, which are often the ones most likely to be reclaimed.

Question 17Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An order workflow calls several services, must retry failed steps with backoff, must wait up to 24 hours for a payment confirmation, and must compensate earlier steps when a later step fails. Which service should orchestrate this?

  1. Amazon EventBridge with scheduled rules
  2. Amazon SNS with an SQS fanout
  3. AWS Lambda with recursive invocations
  4. AWS Step Functions Correct answer.

Explanation

Step Functions expresses multi-step workflows as a state machine with built-in retry, catch, wait, and compensation paths, and it can pause for long durations without holding compute. Chaining Lambda invocations recursively embeds the orchestration logic in code and is difficult to observe and bound.

Question 18Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company wants to route events from many AWS services to different targets based on attributes in the event payload, without writing routing code. Which combination should they use?

  1. One SQS queue per service with a polling consumer
  2. EventBridge rules with event patterns and multiple targets Correct answer.
  3. A Kinesis data stream with a Lambda consumer
  4. A single SNS topic with one Lambda subscriber

Explanation

EventBridge rules match incoming events against event patterns and fan out matching events to one or more targets, so routing is declared rather than coded. SQS, SNS, and Kinesis deliver messages but leave payload-based routing to custom consumer logic.

Question 19Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An application uses ElastiCache for Redis as a session store and cannot lose sessions when a node fails. Which configuration should be used?

  1. A Memcached cluster with Auto Discovery enabled for the clients
  2. A single Redis node with AOF persistence enabled for durability
  3. Redis with replicas and Multi-AZ failover Correct answer.
  4. Redis cluster mode disabled with a single shard

Explanation

Replication plus Multi-AZ automatic failover lets ElastiCache promote a replica in another Availability Zone when the primary fails, keeping the session store available. A single node, with or without persistence, is a single point of failure during node replacement.

Question 20Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A compliance team must retain an RDS for MySQL database backup for three years, while automated backups are configured with a 35-day retention period. What should be done?

  1. Take manual DB snapshots and retain them, optionally copied to another Region Correct answer.
  2. Rely on the automated backups, which are retained indefinitely once created
  3. Set the automated backup retention period to three years
  4. Extend the backup window of the DB instance so backups age more slowly

Explanation

Automated backup retention is capped at 35 days, after which the backups are deleted; manual snapshots persist until explicitly deleted and can be copied cross-Region for durability. Longer retention therefore requires manual snapshots or a managed backup plan.

Question 21Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company runs an active-active application in two Regions backed by DynamoDB global tables. Two users update the same item in different Regions at nearly the same moment. How does DynamoDB resolve the conflict?

  1. Both writes are stored and the application must merge them
  2. The item is locked until both Regions agree
  3. Last-writer-wins based on the item's timestamp Correct answer.
  4. The write to the secondary Region is rejected

Explanation

Global tables use a last-writer-wins reconciliation in which the write with the later timestamp is applied and the earlier write is discarded, giving eventual consistency across replicas. Applications that need to preserve both updates must model them as separate items or use conditional writes to avoid the conflict.

Question 22Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company needs a disaster recovery strategy with an RTO measured in minutes. The recovery Region must always run a scaled-down but fully functional copy of the production stack that can be scaled up quickly. Which strategy is this?

  1. Multi-site active/active
  2. Warm standby Correct answer.
  3. Backup and restore
  4. Pilot light

Explanation

Warm standby keeps a complete but reduced-capacity copy of the workload running in the recovery Region, so recovery consists mainly of scaling up rather than provisioning from scratch. Pilot light keeps only core components such as replicated data running and requires the rest of the stack to be created during recovery.

Question 23Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An application scales EC2 consumers from SQS queue depth. Scaling is erratic: it over-scales on short bursts and under-scales under sustained load. Which metric produces the most stable target tracking?

  1. The number of empty receives recorded by the consumers
  2. Backlog per instance, computed as messages divided by running instances Correct answer.
  3. A fixed schedule that matches historical peak hours
  4. The raw ApproximateNumberOfMessagesVisible metric

Explanation

Dividing the visible message count by the number of running consumers normalizes the backlog to per-instance work, which is what target tracking needs to choose a coherent capacity. The raw queue depth grows with total traffic regardless of how many consumers exist, so it produces unstable scaling decisions.

Question 24Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A distributed database must keep each replica group on independent hardware so that a single hardware failure affects at most one group. Which placement strategy should be used?

  1. A single partition placement group confined to one Availability Zone
  2. A partition placement group across Availability Zones Correct answer.
  3. A spread placement group with one instance per partition
  4. A cluster placement group spread across three Availability Zones

Explanation

Partition placement groups divide instances into logical partitions that each reside on distinct racks, so a rack failure impacts only the instances in that partition, and partitions can be distributed across Availability Zones. Cluster placement groups pack instances closely for low latency and therefore share failure domains.

Question 25Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A Lambda-backed API opens many short-lived connections to an RDS for PostgreSQL instance and exhausts the database connection limit during traffic spikes. Which change addresses the problem with the least application change?

  1. Place an RDS Proxy between the function and the database Correct answer.
  2. Increase the function timeout and retry failed connections with backoff
  3. Raise the function's reserved concurrency so that more invocations can run at once
  4. Increase the DB instance class so that more memory is available for connection overhead

Explanation

RDS Proxy pools and reuses database connections so that many concurrent Lambda invocations share a small number of physical connections, and it also handles failover faster. The function still connects to a proxy endpoint, so no query logic changes are required.

Question 26Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An Auto Scaling group is configured with the EC2 health check type. Instances remain running and pass EC2 status checks, but the application process stops responding to requests. The group does not replace the affected instances. Which change makes the group replace instances whose application is unhealthy?

  1. Use Elastic Load Balancing health checks. Correct answer.
  2. Enable detailed monitoring for the instances in the group.
  3. Attach a lifecycle hook to the group for instance launch.
  4. Increase the group's health check grace period.

Explanation

When an Auto Scaling group uses Elastic Load Balancing health checks, it considers an instance unhealthy when the load balancer reports it as unhealthy. EC2 health checks only reflect instance and hypervisor status, so a running instance with a failed application can still appear healthy.

Question 27Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A standard SQS queue receives messages that a consumer cannot process because of malformed content. The consumer deletes a message only after successful processing, so the same message is received repeatedly. Which feature should be configured to move such messages out of the main queue after several failed receives?

  1. A visibility timeout longer than the processing time.
  2. A delay queue with a per-message timer.
  3. Long polling with a longer wait time.
  4. A dead-letter queue with a redrive policy. Correct answer.

Explanation

A redrive policy specifies a maximum receive count, and after a message exceeds that count SQS moves it to the configured dead-letter queue. This isolates poison messages so they do not circulate in the main queue indefinitely.

Question 28Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A VPC has a single NAT gateway in one Availability Zone. Private subnets in three Availability Zones route 0.0.0.0/0 through that NAT gateway. A failure of the Availability Zone containing the NAT gateway stops outbound internet access for all private subnets. Which design change improves fault tolerance with the least operational overhead?

  1. Replace the NAT gateway with a single NAT instance in the same subnet.
  2. Add a second elastic IP address to the existing NAT gateway.
  3. Create one NAT gateway per Availability Zone with local routes. Correct answer.
  4. Enable VPC Flow Logs on the private subnets.

Explanation

A NAT gateway is scoped to a single Availability Zone, so a single NAT gateway remains a zonal single point of failure. Deploying a NAT gateway in each Availability Zone and routing each subnet through the local NAT gateway removes that dependency.

Question 29Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A CloudFront distribution serves static content from a primary Amazon S3 bucket and must automatically serve from a secondary bucket when the primary returns HTTP 5xx errors. Which CloudFront feature provides this behavior?

  1. Lambda@Edge origin-request triggers.
  2. Field-level encryption for the origin.
  3. Cache behaviors ordered by path pattern.
  4. An origin group with failover criteria. Correct answer.

Explanation

CloudFront origin groups define a primary origin and a secondary origin, and failover occurs based on configured HTTP status codes. This lets CloudFront retry requests against the secondary origin when the primary fails.

Question 30Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A DynamoDB table uses provisioned capacity and experiences unpredictable, short traffic spikes that cause ProvisionedThroughputExceededException errors. The team does not want to manage capacity or build custom scaling. Which change best addresses the problem?

  1. Create a global secondary index for each access pattern.
  2. Switch the table to on-demand capacity mode. Correct answer.
  3. Add a DynamoDB Accelerator (DAX) cluster in front of the table.
  4. Increase the table's provisioned read and write capacity permanently.

Explanation

On-demand capacity mode automatically accommodates traffic without requiring capacity planning or custom scaling policies. It is suited to unpredictable workloads that would otherwise throttle under provisioned capacity.

Question 31Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company wants Amazon S3 Cross-Region Replication to copy objects to a bucket in another Region for disaster recovery. What must be enabled on both the source and destination buckets for replication to work?

  1. Versioning on both buckets. Correct answer.
  2. Transfer Acceleration on both buckets.
  3. Object Lock in compliance mode on both buckets.
  4. MFA Delete on the source bucket.

Explanation

S3 replication requires versioning to be enabled on the source bucket and on the destination bucket. Replication operates on versioned objects and preserves object versions at the destination.

Question 32Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An Amazon ECS service runs many tasks on EC2 instances spread across three Availability Zones. The team wants tasks distributed as evenly as possible across the Availability Zones and instances. Which task placement strategy should be specified in the service definition?

  1. distinctInstance
  2. spread Correct answer.
  3. random
  4. binpack

Explanation

The spread placement strategy distributes tasks evenly across a specified attribute such as Availability Zone or instance ID. This helps balance tasks across the available infrastructure and reduces concentration in one zone.

Question 33Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A consumer application takes up to two minutes to process each message. The queue's visibility timeout is 30 seconds. Other consumers receive and process the same messages again while the first consumer is still working. Which change resolves the duplicate processing?

  1. Add a delay queue with a 30-second delay.
  2. Configure a dead-letter queue for repeatedly failing messages.
  3. Enable long polling on the queue.
  4. Set the visibility timeout above the maximum processing time. Correct answer.

Explanation

The visibility timeout must be longer than the maximum time needed to process and delete a message. Otherwise the message becomes visible again before the first consumer finishes, allowing another consumer to receive it.

Question 34Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An application must deliver each order event to three independent downstream services. Each service needs its own durable buffer, and a failure in one service must not affect the others. Which architecture meets these requirements?

  1. Use an SNS topic with one SQS queue per service. Correct answer.
  2. Publish the event to one Amazon SQS queue that all three services poll.
  3. Send the event directly to each service's HTTP endpoint from the producer.
  4. Write the event to one Amazon S3 object that all three services read.

Explanation

An SNS topic can fan out each event to multiple SQS queues, giving each consumer an independent durable buffer. A failure or slowdown in one consumer does not prevent the others from receiving their own copies.

Question 35Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A relational database must remain available when its primary Availability Zone fails, and the application must continue using the same endpoint. Which RDS configuration provides automatic failover to a standby in another Availability Zone?

  1. A read replica in a different Availability Zone.
  2. A manual snapshot restored in another Availability Zone.
  3. A Multi-AZ DB instance deployment. Correct answer.
  4. A cross-Region read replica.

Explanation

An RDS Multi-AZ DB instance deployment maintains a synchronous standby in another Availability Zone and automatically fails over the DNS endpoint. The application reconnects to the same endpoint after failover.

Question 36Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

Applications on Amazon EC2 instances in three Availability Zones need a shared POSIX-compliant file system. The file system must remain accessible if any one Availability Zone fails. Which configuration should be used?

  1. Instance store volumes replicated by the application across instances.
  2. An Amazon S3 bucket mounted on each instance with a third-party client.
  3. An Amazon EFS file system with a mount target in each Availability Zone. Correct answer.
  4. An Amazon EBS volume with Multi-Attach enabled in one Availability Zone.

Explanation

Amazon EFS is a regional file system, and creating a mount target in each Availability Zone lets instances connect through a local mount target. If one Availability Zone fails, instances in other zones can continue using their own mount targets.

Question 37Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An Auto Scaling group takes about 12 minutes to launch and fully configure new instances, which causes slow responses during sudden scale-out. The team wants a pool of pre-initialized instances that can be brought into service quickly. Which feature should be used?

  1. Predictive scaling with a forecast.
  2. An Auto Scaling group warm pool. Correct answer.
  3. Scheduled scaling for predictable peaks.
  4. A lifecycle hook on instance launch.

Explanation

A warm pool maintains pre-initialized instances that can be added to the Auto Scaling group quickly instead of launching and configuring from scratch. This reduces scale-out latency for workloads with long bootstrap times.

Question 38Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

A company uses Route 53 Application Recovery Controller routing controls to shift traffic between two active-passive cells. Operators must be prevented from disabling both cells at the same time during a failover. Which ARC feature enforces this constraint?

  1. Latency-based routing with health checks.
  2. Health checks with calculated status.
  3. Route 53 failover routing policies.
  4. Routing control safety rules. Correct answer.

Explanation

ARC routing control safety rules assert conditions about routing control states, such as requiring at least one routing control to remain on. This prevents an operator action from turning off all traffic paths at once.

Question 39Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An Amazon Kinesis Data Streams application writes records with a partition key derived from customer ID. One very large customer generates most of the traffic, and the stream throttles writes even though total shard capacity appears sufficient. Which change addresses the hot-shard problem?

  1. Add a random suffix to the partition key. Correct answer.
  2. Increase the stream's retention period.
  3. Increase the number of consumers reading from the stream.
  4. Enable enhanced fan-out for the consumers.

Explanation

Kinesis uses the partition key to map each record to a shard, so a dominant key can overload one shard while others remain underused. Adding a random suffix spreads that key's records across more shards, provided per-customer ordering is not required.

Question 40Designing and deploying dynamically scalable, highly available, fault-tolerant, and reliable applications on AWS

An Amazon ECS service on Fargate deploys a new task definition. New tasks start but fail their load balancer health checks, and the deployment continues replacing tasks without stabilizing. Which feature should be enabled so the service automatically stops and rolls back a failed deployment?

  1. The deployment circuit breaker with rollback. Correct answer.
  2. A capacity provider strategy with Fargate.
  3. Container health checks defined in the task definition.
  4. Service Auto Scaling with target tracking.

Explanation

The ECS deployment circuit breaker detects when a deployment cannot reach a steady state and can automatically roll back to the previous task definition. Enabling rollback prevents a failing deployment from continuing to replace tasks indefinitely.

Question 41Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A company needs to store and serve millions of static images and videos for a global web application. The objects are accessed over HTTPS and must be durably stored across multiple Availability Zones without managing servers. Which service should the company use?

  1. Amazon EBS volumes attached to web servers
  2. Amazon EFS mounted on web servers
  3. Amazon S3 with CloudFront distribution Correct answer.
  4. Amazon FSx for Windows File Server

Explanation

Amazon S3 is object storage that automatically stores data redundantly across multiple Availability Zones within a Region, and CloudFront caches objects at edge locations for global low-latency delivery. EBS volumes are block storage tied to a single Availability Zone, and file systems are not designed for global static website delivery.

Question 42Selecting appropriate AWS services to design and deploy an application based on a set of requirements

An application produces messages that must be processed asynchronously by a fleet of workers. Each message must be processed by one worker, and failed messages should be retried after a delay. Which service should the company use?

  1. Amazon SNS topic with an email subscription
  2. Amazon SQS standard queue with a dead-letter queue Correct answer.
  3. Amazon EventBridge rule with a Lambda target
  4. Amazon Kinesis Data Streams with multiple consumers

Explanation

Amazon SQS is a pull-based queue in which each message is delivered to one consumer, and the visibility timeout controls when an unacknowledged message becomes available again. A dead-letter queue captures messages that exceed the configured receive limit.

Question 43Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A mobile app needs to authenticate users with email and password, support social identity providers, and issue JSON Web Tokens that authorize API calls. Which service should the company use?

  1. Amazon Cognito identity pools
  2. AWS IAM Identity Center
  3. Amazon Cognito user pools Correct answer.
  4. AWS Secrets Manager

Explanation

Cognito user pools provide a managed user directory, federation with social identity providers, and issuance of OIDC/JWT tokens for API authorization. Identity pools instead exchange identities for temporary AWS credentials.

Question 44Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A company hosts a static website in an S3 bucket that must not be publicly accessible. Users worldwide should access the site with low latency over HTTPS. Which configuration should the company use?

  1. S3 Transfer Acceleration endpoint
  2. S3 website endpoint with public read access
  3. CloudFront with an S3 origin access control and bucket policy Correct answer.
  4. CloudFront with an S3 origin access identity and public bucket

Explanation

CloudFront with an origin access control lets CloudFront access a private S3 bucket through a bucket policy while serving content over HTTPS from edge locations. OAC is the current mechanism for securing S3 origins.

Question 45Selecting appropriate AWS services to design and deploy an application based on a set of requirements

An application retrieves a database password at runtime. Security policy requires automatic rotation every 30 days and auditing of secret access. Which solution should the company use?

  1. Store the password in AWS Secrets Manager with automatic rotation Correct answer.
  2. Store the password in an encrypted S3 object
  3. Store the password in AWS Systems Manager Parameter Store SecureString
  4. Embed the password in the application source code

Explanation

Secrets Manager is designed for storing credentials and supports built-in automatic rotation for supported databases such as Amazon RDS. It also integrates with AWS CloudTrail for auditing secret access.

Question 46Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A gaming application needs a real-time leaderboard that supports ranking millions of players with low-latency reads and writes. Which service should the company use?

  1. Amazon ElastiCache for Redis with sorted sets Correct answer.
  2. Amazon RDS for PostgreSQL with a read replica
  3. Amazon DynamoDB with a global secondary index
  4. Amazon ElastiCache for Memcached

Explanation

Redis sorted sets provide efficient ordered ranking with low-latency updates and range queries, which fits real-time leaderboard workloads. ElastiCache for Redis is a managed in-memory service for this pattern.

Question 47Selecting appropriate AWS services to design and deploy an application based on a set of requirements

Analysts need to run ad hoc SQL queries against large CSV and Parquet files stored in S3 without loading the data into a database. Which service should the company use?

  1. Amazon Athena with a Glue table Correct answer.
  2. Amazon EMR with a custom MapReduce job
  3. Amazon Redshift with COPY from S3
  4. Amazon RDS for MySQL with a stored procedure

Explanation

Athena is a serverless query service that runs SQL directly on data in S3, using the AWS Glue Data Catalog to define table schemas. It does not require loading the data into a database cluster.

Question 48Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A company stores millions of objects in S3 with unpredictable and changing access patterns. It wants automatic cost optimization without retrieval fees for infrequent access. Which storage class should the company use?

  1. S3 Standard for all objects
  2. S3 Intelligent-Tiering Correct answer.
  3. S3 Glacier Deep Archive for all objects
  4. S3 One Zone-IA for all objects

Explanation

S3 Intelligent-Tiering monitors access patterns and moves objects between access tiers automatically, with no retrieval fees for the infrequent access tiers. It is intended for data with unknown or changing access patterns.

Question 49Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A company exposes a REST API to external partners. It needs per-partner API keys, request throttling, and usage plans with quotas. Which service should the company use?

  1. Application Load Balancer with host-based routing
  2. Network Load Balancer with TLS termination
  3. API Gateway REST API with usage plans Correct answer.
  4. AWS AppSync GraphQL API with API keys

Explanation

API Gateway provides API keys, usage plans, throttling, and quotas for REST APIs, which are designed for partner-specific access control and rate limiting. Load balancers route traffic but do not provide API key usage plans.

Question 50Selecting appropriate AWS services to design and deploy an application based on a set of requirements

A company wants to stream application logs into S3 in near real time with automatic buffering, compression, and format conversion, without writing consumer code. Which service should the company use?

  1. Amazon Kinesis Data Streams with a custom consumer
  2. Amazon SQS FIFO queue with a Lambda consumer
  3. Amazon MSK with a Kafka Connect S3 sink
  4. Amazon Data Firehose delivery stream Correct answer.

Explanation

Amazon Data Firehose is a managed delivery service that buffers, compresses, transforms, and delivers streaming data to destinations such as S3. It eliminates the need to write and operate consumer applications.