Design for New Solutions

58. A deployment strategy must allow an immediate return to the previous version if the new version misbehaves, without waiting for instances to be replaced. Which solution meets these requirements?

1. A Deploy with a blue/green strategy so the previous environment remains running and traffic can be shifted back.
2. B Deploy with a rolling strategy so instances are replaced in batches.
3. C Deploy with an all-at-once strategy during a maintenance window.
4. D Deploy with a rolling strategy and an additional batch to preserve capacity.

59. An architect must reduce the operational burden of patching and provisioning for a new three-tier application. Which combination of choices meets these requirements? (Select TWO.)

1. A Run the application tier on EC2 instances in an Auto Scaling group with a golden AMI pipeline.
2. B Run the application tier on EC2 instances with a Systems Manager patch baseline.
3. C Use Amazon Aurora Serverless v2 for the database so capacity and patching are managed.
4. D Run the application tier on AWS Fargate so no instances are patched or provisioned.
5. E Run the database on EC2 instances with automated snapshot scheduling.

60. A business continuity plan must be exercised regularly so recovery procedures are known to work. Which solution meets these requirements?

1. A Review the recovery runbook with the response team each quarter.
2. B Confirm that cross-Region backup copies complete successfully each night.
3. C Run periodic failover exercises in which the standby serves production traffic, and record the outcome.
4. D Verify that the standby Region resources exist and match the primary.

61. A relational database must be recoverable in a second Region with a recovery point objective under one minute. Which solution meets these requirements?

1. A Copy automated snapshots to the second Region every hour.
2. B Deploy an Amazon Aurora global database with a secondary Region and promote it during failover.
3. C Deploy a Multi-AZ database with three read replicas in the primary Region.
4. D Enable point-in-time recovery and restore into the second Region when needed.

62. An architect must design a backup strategy covering EBS volumes, RDS databases, EFS file systems, and DynamoDB tables under one policy with cross-Region copies. Which solution meets these requirements?

1. A Create Amazon Data Lifecycle Manager policies for each resource type.
2. B Create a Lambda function that invokes each service's snapshot API on a schedule.
3. C Create an AWS Backup plan covering the supported resource types with a cross-Region copy rule.
4. D Create manual snapshots of each resource before major changes.

63. A large public web application must be protected against volumetric DDoS attacks and application-layer exploits, with cost protection against attack-driven scaling. Which combination of steps meets these requirements? (Select TWO.)

1. A Configure a security group that permits inbound traffic only on port 443.
2. B Associate an AWS WAF web ACL with managed rule groups and rate-based rules.
3. C Enable Amazon GuardDuty and review its findings daily.
4. D Subscribe to AWS Shield Advanced and add the application's resources as protected resources.
5. E Enable AWS Shield Standard, which is included for all customers.

64. An architect must specify how a new workload's roles are scoped so no identity holds more permission than its function requires. Which approach meets these requirements?

1. A Define one role per person so permissions can be traced to an individual.
2. B Define roles with wildcard actions and add explicit denies for sensitive operations.
3. C Define one role for the workload with a managed policy granting broad access to the services involved.
4. D Define a role per function with policies naming only the required actions and resource ARNs, and attach a permissions boundary where roles create roles.

65. A new application must remain available when an Availability Zone fails, and peak load requires 12 instances. Which solution meets these requirements at the lowest instance count?

1. A Run 18 instances distributed evenly across three Availability Zones.
2. B Run 12 instances in one Availability Zone with a standby Amazon Machine Image.
3. C Run 24 instances distributed evenly across two Availability Zones.
4. D Run 12 instances distributed evenly across three Availability Zones.

66. An architect must decouple a synchronous order service from a fulfilment service that is occasionally unavailable. Which solution meets these requirements?

1. A Increase the fulfilment service's instance count so it is less likely to be unavailable.
2. B Run both services in the same Auto Scaling group so they scale together.
3. C Write order events to an Amazon SQS queue that the fulfilment service consumes independently.
4. D Call the fulfilment service synchronously with a short timeout and retries.

67. Users worldwide must be routed to the Regional endpoint nearest them, and an unhealthy Region must be removed from rotation. Which solution meets these requirements?

1. A Create Route 53 latency records for each Regional endpoint and attach a health check to each record.
2. B Create Route 53 weighted records split evenly between the Regions.
3. C Create Route 53 simple records returning all Regional endpoints.
4. D Create Route 53 geolocation records mapping each country to a Region.

68. A read-heavy workload must serve a small set of frequently requested items with sub-millisecond latency. Which solution meets these requirements?

1. A Increase the database's provisioned IOPS so storage responds faster.
2. B Add read replicas to the database and direct the repeated reads to them.
3. C Place an in-memory cache in front of the data store and serve the repeated reads from it.
4. D Increase the database instance class so queries complete more quickly.

69. An architect must select a database for a workload requiring millisecond key-value lookups at unpredictable scale with no capacity planning. Which solution meets these requirements?

1. A Use Amazon Aurora with auto scaling read replicas.
2. B Use Amazon RDS with a read replica and connection pooling.
3. C Use Amazon Redshift with a sort key on the lookup column.
4. D Use Amazon DynamoDB in on-demand capacity mode.

70. A new workload's data will be accessed heavily for 30 days, occasionally for 90 more, and rarely afterwards, with a seven-year retention requirement. Which solution meets these requirements at the lowest cost?

1. A Apply an S3 Lifecycle configuration expiring the objects after 120 days.
2. B Store all objects in S3 Standard for the full retention period.
3. C Store all objects in S3 Intelligent-Tiering for the full retention period.
4. D Apply an S3 Lifecycle configuration transitioning to Standard-Infrequent Access at 30 days and to Glacier Deep Archive at 120 days.

71. An architect must reduce the data transfer cost of a new application serving large media files to internet users. Which solution meets these requirements?

1. A Store the media in S3 Glacier Instant Retrieval to reduce storage cost.
2. B Move the origin to a Region with lower compute pricing.
3. C Serve the media through Amazon CloudFront so repeat requests are cached and egress is priced lower.
4. D Increase the origin instance size so files are served more quickly.

72. An architect must select a deployment approach for a new workload where a failed release must be reversible within seconds. Which solution meets these requirements?

1. A Deploy with a rolling strategy so instances are replaced in small batches.
2. B Deploy with an all-at-once strategy inside a maintenance window.
3. C Deploy with a blue/green strategy so the previous environment remains available for an immediate traffic shift back.
4. D Deploy with a rolling strategy and an additional batch to preserve capacity.

73. An architect must reduce the operational burden of a new workload's configuration management across hundreds of instances. Which solution meets these requirements?

1. A Configure each instance during launch with a user data script.
2. B Document the required configuration in a runbook for operators to apply.
3. C Create a golden AMI and rebuild instances whenever configuration changes.
4. D Manage the desired configuration with AWS Systems Manager State Manager so drift is corrected automatically.

74. A new workload must adopt managed services wherever possible to reduce undifferentiated operational work. Which combination of choices supports this? (Select TWO.)

1. A Run the database on EC2 instances with automated snapshots configured.
2. B Use Amazon Aurora Serverless v2 rather than a self-managed database on EC2.
3. C Run the application tier on AWS Fargate rather than self-managed instances.
4. D Run both tiers on EC2 instances in a single Auto Scaling group.
5. E Run the application tier on EC2 instances with a patch baseline applied.

75. A new workload's database must fail over to a second Region with a recovery point measured in seconds. Which solution meets these requirements?

1. A Deploy an Amazon Aurora global database with a secondary Region that can be promoted.
2. B Copy automated snapshots to the second Region every fifteen minutes.
3. C Configure cross-Region read replicas and promote one manually during failover.
4. D Deploy a Multi-AZ database with synchronous standby in the primary Region.

76. An architect must design DNS failover so users are directed away from a Region whose application has failed while its infrastructure remains reachable. Which solution meets these requirements?

1. A Attach a health check evaluating each Region's load balancer reachability.
2. B Reduce the record's time to live so clients refresh sooner.
3. C Configure weighted routing split evenly between the Regions.
4. D Attach a health check evaluating an application endpoint to each Regional record so a failed application removes the record.

77. An architect must ensure that a disaster recovery plan for a new workload is exercised without disrupting production. Which approach is appropriate?

1. A Confirm each night that replication to the standby Region has completed.
2. B Fail over to the standby environment during a scheduled window and serve real traffic from it, then fail back.
3. C Restore the standby from backup into an isolated account and confirm it starts.
4. D Review the recovery runbook with the operations team each quarter.

78. An architect must design centralized monitoring so a failure in any component of a new multi-Region workload is detected proactively. Which solution meets these requirements?

1. A Alarm on CPU utilization in each Region.
2. B Create a dashboard in each Region for operators to review each morning.
3. C Aggregate metrics and alarms into a central account and alarm on the absence of expected success signals as well as on errors.
4. D Enable detailed monitoring on every instance in both Regions.

79. A new public API must authenticate end users and authorize requests without the application handling credentials. Which solution meets these requirements?

1. A Issue each user an IAM access key and validate it in the application.
2. B Require a shared API key sent in a request header.
3. C Place the API behind a load balancer restricted by source IP address.
4. D Authenticate users with Amazon Cognito and authorize API Gateway requests with the resulting tokens.

80. An architect must specify network flows for a new three-tier workload so each tier accepts traffic only from the tier above it. Which solution meets these requirements?

1. A Apply network ACLs permitting traffic between the subnet ranges.
2. B Reference each tier's subnet CIDR range as the source in the next tier's inbound rules.
3. C Reference each tier's security group as the source in the next tier's inbound rules.
4. D Place each tier in a separate VPC and peer them.

81. An architect must define a patch management strategy for a new workload that must remain compliant with an organizational standard. Which solution meets these requirements?

1. A Instruct operators to patch instances when vulnerabilities are announced.
2. B Define a Systems Manager patch baseline, apply it through a maintenance window, and report compliance per instance.
3. C Configure each instance to install all available updates nightly.
4. D Rebuild every instance from a new AMI each week.

82. A new application must mitigate large-scale attacks against its public endpoint while keeping legitimate traffic served. Which combination of steps meets these requirements? (Select TWO.)

1. A Associate an AWS WAF web ACL with managed rule groups and rate-based rules.
2. B Increase the instance size of the application tier.
3. C Enable detailed CloudWatch monitoring on the endpoint.
4. D Restrict the endpoint's security group to port 443 only.
5. E Subscribe to AWS Shield Advanced and register the endpoint as a protected resource.

83. An architect must design a new workload so a failure in one component cannot exhaust the resources of another. Which pattern applies?

1. A Retry with exponential backoff between the components.
2. B Bulkhead isolation, where each component has its own resource pool.
3. C A shared connection pool sized for peak demand.
4. D A single Auto Scaling group containing both components.

84. A new workload depends on a downstream service that occasionally becomes unresponsive, causing requests to queue until the workload exhausts its threads. Which pattern applies?

1. A A circuit breaker that stops calling the downstream service after repeated failures and fails fast.
2. B Retrying each failed request immediately until it succeeds.
3. C Increasing the thread pool size so more requests can wait.
4. D Increasing the request timeout so slow responses complete.

85. An architect must ensure a new workload's Auto Scaling group replaces instances that are running but not serving requests correctly. Which configuration meets these requirements?

1. A Increase the group's desired capacity so healthy instances outnumber failed ones.
2. B Configure the group to use EC2 health checks so unreachable instances are replaced.
3. C Reduce the group's health check grace period so failures are detected sooner.
4. D Configure the group to use ELB health checks so application-level failures trigger replacement.

86. A new workload must stay within a service quota that its expected growth will approach within six months. Which approach is appropriate?

1. A Design the workload to fail gracefully when the quota is reached.
2. B Monitor the application's error rate and request an increase when errors appear.
3. C Alarm on the Service Quotas usage metric at a proportion of the limit and request an increase before the limit is reached.
4. D Distribute the workload across additional Regions to avoid the quota.

87. An architect must choose a routing policy so users are served from the Region nearest them, with an unhealthy Region removed automatically. Which solution meets these requirements?

1. A Route 53 geolocation records mapping each country to a Region.
2. B Route 53 weighted records split in proportion to Regional capacity.
3. C Route 53 multivalue answer records returning all Regional endpoints.
4. D Route 53 latency records with a health check attached to each record.

88. A new workload writes large objects and reads them repeatedly from many Regions. Which solution meets these requirements?

1. A Store the objects in Amazon S3 and distribute them through Amazon CloudFront.
2. B Store the objects on Amazon EBS volumes attached to instances in each Region.
3. C Store the objects in Amazon EFS and mount it from each Region.
4. D Store the objects in Amazon S3 and replicate the bucket to every Region.

89. A new analytics workload issues complex joins over 30 TB with high concurrency. Which solution meets these requirements?

1. A Load the data into Amazon Redshift and define distribution and sort keys matching the query patterns.
2. B Load the data into Amazon RDS on a memory-optimized instance class.
3. C Query the data in place with Amazon Athena over partitioned files.
4. D Load the data into Amazon DynamoDB with a global secondary index per query.

90. An architect must design a new workload to absorb a spike that arrives faster than the downstream system can process. Which pattern applies?

1. A Scale the downstream system to the size of the largest expected spike.
2. B Reject requests that exceed the downstream system's capacity.
3. C Increase the request timeout so slow processing completes.
4. D Buffer the requests in a queue so the consumer processes them at a sustainable rate.

91. A new workload requires a purpose-built database for highly connected data queried by relationship depth. Which solution meets these requirements?

1. A Amazon RDS for PostgreSQL
2. B Amazon DynamoDB
3. C Amazon Neptune
4. D Amazon Redshift

92. An architect must model the data transfer cost of a new multi-Region design before it is built. Which approach is appropriate?

1. A Assume data transfer is a negligible proportion of total cost.
2. B Measure the transfer cost after the workload has been running for a month.
3. C Price the design using only its compute and storage components.
4. D Estimate the volume crossing each boundary and price it with the AWS Pricing Calculator, since inter-Region and internet egress are charged differently.

93. A new workload has a predictable steady baseline and unpredictable bursts. Which purchasing approach is appropriate?

1. A Cover the entire expected peak with a Savings Plan.
2. B Cover the baseline with a Savings Plan and serve the bursts with On-Demand capacity.
3. C Serve all capacity with Spot Instances to minimise cost.
4. D Serve all capacity with On-Demand to retain flexibility.

94. An architect must plan an upgrade path for a workload to adopt a new major version of a managed service with breaking changes. Which approach is appropriate?

1. A Delay the upgrade until the old version is deprecated.
2. B Provision the new version alongside the existing one, validate the workload against it, and migrate with a defined rollback.
3. C Upgrade the existing service in place during a maintenance window.
4. D Upgrade in place and fix breaking changes as they appear.

95. A new workload must be deployed across many accounts with a consistent configuration that account owners cannot alter. Which solution meets these requirements?

1. A Deploy with CloudFormation StackSets using service-managed permissions and apply a service control policy preventing modification of the deployed resources.
2. B Deploy with StackSets and trust account owners not to modify the resources.
3. C Deploy manually in each account following a runbook.
4. D Send each account owner the template to deploy.

96. A workload's change management process must ensure that every infrastructure change is reviewed and reversible. Which approach is appropriate?

1. A Manage infrastructure as code in version control with reviewed merges, and deploy through a pipeline that supports rollback to the previous version.
2. B Allow senior engineers to make changes directly.
3. C Apply changes in the console and document them in a ticket.
4. D Review changes quarterly in an audit.

97. A workload's recovery point objective is one hour, and its database is replicated asynchronously to a second Region. Which additional consideration applies?

1. A Asynchronous replication guarantees the objective by design.
2. B The replication should be switched to synchronous for safety.
3. C The objective should be relaxed to match the replication method.
4. D Replication lag must be monitored and alarmed so a lag approaching one hour is detected before an incident exposes it.

98. An architect must design DNS so that failover to a second Region does not depend on the primary Region being available to make the change. Which solution meets these requirements?

1. A Use Route 53 health checks with failover records, since Route 53 is a global service that evaluates health independently of either Region.
2. B Run a script in the primary Region that updates DNS on failure.
3. C Use a load balancer in the primary Region to route to the secondary.
4. D Have an operator update DNS manually during an incident.

99. A disaster recovery test must be performed without affecting production users. Which approach is appropriate?

1. A Fail over a copy of production traffic or a synthetic workload to the standby, validate it, and fail back without touching live users.
2. B Test only the backup restore step in isolation.
3. C Skip testing to avoid any risk to production.
4. D Fail over all production traffic during business hours.

100. A multi-Region application's data must be readable and writable in both Regions with conflict resolution. Which solution meets these requirements?

1. A Use a single-Region database with read replicas in the second Region.
2. B Copy the database between Regions nightly.
3. C Use a database with multi-Region active-active support and built-in conflict resolution, such as DynamoDB global tables.
4. D Use an Aurora global database, which permits writes in one Region only.

101. An organization must confirm that its standby Region has sufficient capacity and quota to absorb production load before a failover is needed. Which solution meets these requirements?

1. A Request quota increases at the time of failover.
2. B Check the quotas manually once when the standby is built.
3. C Assume the standby Region's default quotas suffice.
4. D Use Amazon Application Recovery Controller readiness checks to continuously verify capacity, quotas, and configuration in the standby.

102. Centralized monitoring must detect a failure in a Region's application before the Region's own monitoring can report it. Which approach is appropriate?

1. A Rely on each Region's CloudWatch alarms.
2. B Run synthetic canaries from a separate Region against each Region's endpoints so detection does not depend on the monitored Region.
3. C Check each Region's status page.
4. D Ask users to report outages.

103. A new workload must call AWS services from EC2 instances without the instances holding any credentials. Which solution meets these requirements?

1. A Retrieve access keys from Secrets Manager at startup.
2. B Store access keys in the instance's user data.
3. C Attach an instance profile so the instances receive temporary credentials automatically.
4. D Store access keys in a configuration file on each instance.

104. A new application's secrets must be rotated automatically, and the application must not need redeployment when they rotate. Which solution meets these requirements?

1. A Store the secrets in AWS Secrets Manager with rotation enabled and have the application retrieve them at use time.
2. B Hardcode the secrets and rotate them quarterly.
3. C Store the secrets in a configuration file updated by a script.
4. D Store the secrets in environment variables and redeploy on rotation.

105. A new workload must reach Amazon S3 and DynamoDB from private subnets without a NAT gateway. Which solution meets these requirements?

1. A Assign public IP addresses to the instances.
2. B Create gateway VPC endpoints for both services and add routes to the private subnet route tables.
3. C Route through a proxy in a public subnet.
4. D Create interface VPC endpoints for both services.

106. A new web application must be protected against attacks that exploit application logic rather than known signatures. Which combination of steps meets these requirements? (Select TWO.)

1. A Enable Shield Standard.
2. B Implement application-level validation and authorization checks, since WAF cannot understand business logic.
3. C Apply WAF rate-based rules to limit abusive request patterns.
4. D Increase the application's instance size.
5. E Rely on WAF managed rule groups alone.

107. An architect must decide how to apply patches to a new fleet where instances are replaced frequently. Which approach is appropriate?

1. A Skip patching because instances are short-lived.
2. B Bake patches into the machine image on a regular cycle and replace instances from the updated image.
3. C Patch instances manually when vulnerabilities are announced.
4. D Patch each running instance in place on a schedule.

108. A new workload must remain available during an Availability Zone failure without over-provisioning by more than 50 percent. Which configuration meets these requirements?

1. A Deploy in one Availability Zone with a standby image.
2. B Deploy across three Availability Zones with capacity for peak in any two, which over-provisions by 50 percent.
3. C Deploy across two Availability Zones with full peak capacity in each, which over-provisions by 100 percent.
4. D Deploy across three Availability Zones with peak capacity divided equally, which fails if a zone is lost.

109. An application's components communicate through a message queue, and a consumer occasionally fails to process a message. Which configuration prevents a poison message from blocking the queue?

1. A Configure a dead-letter queue with a maximum receive count so a repeatedly failing message is moved aside.
2. B Delete messages that fail on first processing.
3. C Retry each message indefinitely.
4. D Increase the queue's visibility timeout.

110. A new workload must recover from a database failover with minimal application disruption. Which combination of steps meets these requirements? (Select TWO.)

1. A Disable the application's connection timeout.
2. B Hardcode the primary instance's IP address for lowest latency.
3. C Connect through the database's cluster endpoint rather than an instance address, so failover updates the target.
4. D Keep a single long-lived connection open.
5. E Implement connection retry with backoff in the application, since existing connections drop at failover.

111. A new application must fail over between Regions based on the health of the application rather than the infrastructure. Which health check configuration is appropriate?

1. A A Route 53 health check against the load balancer's DNS name.
2. B A CloudWatch alarm on instance CPU.
3. C A Route 53 health check against an application endpoint that exercises a real transaction path.
4. D An EC2 status check.

112. A new workload must scale on a metric that reflects user-facing performance rather than resource utilization. Which configuration is appropriate?

1. A Scheduled scaling at fixed times.
2. B Target tracking on average CPU utilization.
3. C Manual scaling by operators.
4. D Target tracking on request latency or request count per target, which follows user demand directly.

113. A new application must be designed so a failure in the reporting subsystem cannot affect order processing. Which pattern applies?

1. A Run both subsystems in the same process for simplicity.
2. B Decouple the subsystems through asynchronous messaging so each runs and fails independently.
3. C Deploy both subsystems in the same Auto Scaling group.
4. D Call reporting synchronously from order processing with a timeout.

114. A new workload must serve global users with dynamic content that cannot be cached. Which solution meets these requirements?

1. A Deploy in multiple Regions with latency-based routing, or use Global Accelerator to bring traffic onto the AWS network near users.
2. B Deploy in one Region behind CloudFront with caching disabled.
3. C Deploy in one Region and compress responses.
4. D Deploy in one Region on the largest instances.

115. A new database workload has unpredictable traffic with long idle periods. Which solution meets these requirements?

1. A Aurora Serverless v2, which scales capacity with load and down during idle periods.
2. B A provisioned RDS instance with read replicas.
3. C A provisioned Aurora cluster sized for peak.
4. D A DynamoDB table with provisioned capacity.

116. An architect must select storage for a workload requiring shared POSIX file access from many instances with high throughput. Which solution meets these requirements?

1. A Instance store on each instance.
2. B Amazon S3 mounted as a file system.
3. C Amazon EBS attached to each instance.
4. D Amazon EFS or FSx for Lustre, depending on throughput and latency requirements.

117. A new workload's read traffic is dominated by a small set of frequently accessed keys. Which pattern is appropriate?

1. A Add read replicas to the data store.
2. B Place an in-memory cache in front of the data store so hot keys are served without reaching it.
3. C Increase the data store's instance size.
4. D Partition the data store by key range.

118. A new workload must process a very large batch job as quickly as possible with the ability to use spare capacity. Which solution meets these requirements?

1. A Run the job on AWS Batch with a mix of Spot and On-Demand capacity across many instance types.
2. B Run the job on a fixed Auto Scaling group of On-Demand instances.
3. C Run the job on Lambda functions.
4. D Run the job on a single large On-Demand instance.

119. A new workload will run steadily for three years, and the organization wants the largest possible discount with flexibility to change instance family. Which purchasing option meets these requirements?

1. A A three-year Compute Savings Plan, which applies across instance families and Regions.
2. B Three-year Standard Reserved Instances for the initial instance family.
3. C On-Demand capacity with a budget alert.
4. D One-year Convertible Reserved Instances renewed annually.

120. A new workload will generate large volumes of logs that must be retained for seven years but rarely read. Which solution meets these requirements at the lowest cost?

1. A Store the logs in S3 Standard for seven years.
2. B Retain the logs in CloudWatch Logs for seven years.
3. C Store the logs in an EBS volume.
4. D Deliver the logs to S3 with a lifecycle policy transitioning to Glacier Deep Archive after the active period.

121. An architect must design a new workload so that its cost can be tracked per customer. Which approach is appropriate?

1. A Track cost only at the workload level.
2. B Deploy a separate account per customer.
3. C Tag resources per customer where possible, and for shared resources, meter usage per customer and allocate cost proportionally.
4. D Estimate each customer's cost as an equal share.

122. A new workload must be deployed across accounts with each environment's configuration held separately from the template. Which solution meets these requirements?

1. A Deploy manually with the correct values entered each time.
2. B Maintain a separate template per environment.
3. C Hardcode each environment's values and switch templates.
4. D Store environment configuration in Parameter Store or AppConfig and reference it at deployment so one template serves every environment.

123. A new workload's deployment must be validated in a production-like environment before release. Which solution meets these requirements?

1. A Validate in production during a maintenance window.
2. B Validate in a development environment built by hand.
3. C Rely on unit tests only.
4. D Deploy to a staging environment built from the same infrastructure definition as production and run the validation suite there.

124. A new workload must support deploying individual components independently. Which solution meets these requirements?

1. A Define each component's infrastructure in its own stack with explicit interfaces between them.
2. B Deploy all components together on a fixed schedule.
3. C Deploy components manually in dependency order.
4. D Define all components in one stack.

125. An architect must choose a deployment approach for a workload where the database schema change is not backward compatible. Which solution meets these requirements?

1. A Roll back the schema if the application fails.
2. B Deploy the schema and application changes simultaneously.
3. C Deploy the schema change in a backward-compatible sequence of steps so old and new application versions can both run during the rollout.
4. D Take the application offline for the schema change.

126. A new workload's recovery must be tested without affecting the production environment. Which solution meets these requirements?

1. A Recover into an isolated environment from the same recovery points production would use, and validate the result.
2. B Verify the backup jobs complete successfully.
3. C Review the recovery runbook with the team.
4. D Recover over the production environment during a quiet period.

127. A new workload's data must be recoverable after a logical corruption that replicates to the standby. Which solution meets these requirements?

1. A Rely on the cross-Region replica.
2. B Retain point-in-time backups independent of the replication, since replication propagates the corruption.
3. C Add a second replica in another Region.
4. D Increase the replication frequency.

128. A new workload must continue accepting writes during a Regional failure. Which solution meets these requirements?

1. A Deploy a pilot light in the second Region.
2. B Deploy an active-passive design with asynchronous replication.
3. C Deploy a warm standby in the second Region.
4. D Deploy an active-active design with a data store that accepts writes in both Regions and resolves conflicts.

129. A business continuity plan must define what the business does while a workload is unavailable. Which solution meets these requirements?

1. A Define the manual or degraded process the business follows during an outage, alongside the technical recovery plan.
2. B Define only the technical recovery steps.
3. C Document the outage in the incident record afterwards.
4. D Assume the recovery time objective is short enough that no process is needed.

130. An architect must decide the recovery point objective for a new workload. Which solution meets these requirements?

1. A Derive it from what data loss the business can tolerate, then select a replication approach that meets it.
2. B Set it to match the replication method already available.
3. C Set it to the longest interval the technology permits.
4. D Set it to zero for every workload.

131. A new workload must encrypt data with a key that can be revoked instantly across all copies. Which solution meets these requirements?

1. A Rotate the key's material.
2. B Encrypt each copy with a separate key.
3. C Delete the copies individually.
4. D Encrypt with a KMS key and disable or delete its material to render every copy unreadable.

132. A new workload must restrict which principals can decrypt its data independently of who can read the objects. Which solution meets these requirements?

1. A Use SSE-KMS with a customer managed key whose key policy names the permitted principals.
2. B Enable Block Public Access on the bucket.
3. C Use client-side encryption with a shared key.
4. D Use SSE-S3 and restrict the bucket policy.

133. A new workload must ensure that a compromised application cannot read other tenants' data. Which solution meets these requirements?

1. A Rely on the application filtering by tenant identifier.
2. B Encrypt each tenant's data with the same key.
3. C Store each tenant's data in a separate prefix.
4. D Scope the application's credentials per tenant so the permissions it holds at any moment cover only that tenant.

134. A new workload must authenticate service-to-service calls without shared secrets. Which solution meets these requirements?

1. A Use IAM authentication so each service's role identity is verified by the receiving service.
2. B Restrict calls by source IP address.
3. C Use a shared bearer token stored in configuration.
4. D Share an API key between the services.

135. A new workload must ensure that a misconfigured resource policy cannot expose data outside the organization. Which solution meets these requirements?

1. A Apply a service control policy denying policy changes.
2. B Apply a resource control policy at the organization level so access is bounded regardless of the resource policy.
3. C Enable IAM Access Analyzer and archive known findings.
4. D Review resource policies before deployment.

136. A new workload must log every access to sensitive data for later audit. Which solution meets these requirements?

1. A Enable CloudTrail management events.
2. B Enable server access logging on the storage.
3. C Enable CloudTrail data events for the relevant resources before the workload goes live.
4. D Review access patterns quarterly.

137. A new workload must be prevented from being deployed with an overly permissive security group. Which solution meets these requirements?

1. A Review the security groups after deployment.
2. B Create a Config rule reporting permissive groups.
3. C Evaluate the infrastructure definition with policy-as-code in the pipeline and fail the build on a violation.
4. D Restrict who may create security groups.

138. A new workload must degrade gracefully when a non-essential dependency is unavailable. Which solution meets these requirements?

1. A Remove the non-essential dependency from the design.
2. B Retry the non-essential dependency until it responds.
3. C Identify which dependencies are essential and design the application to continue with reduced function when a non-essential one fails.
4. D Treat every dependency as essential and fail the request.

139. A new workload must avoid a retry storm when a downstream service recovers from an outage. Which solution meets these requirements?

1. A Retry immediately when the service recovers.
2. B Apply exponential backoff with jitter so clients do not converge on the same retry instant.
3. C Increase the downstream service's capacity.
4. D Have all clients retry on a fixed schedule.

140. A new workload must continue operating when its configuration store is briefly unavailable. Which solution meets these requirements?

1. A Cache the configuration locally with a defined staleness tolerance so the workload continues during a brief outage.
2. B Fail requests when the configuration store is unavailable.
3. C Retry the configuration store until it responds.
4. D Embed the configuration in the deployment artifact.

141. An architect must verify a new workload meets its availability target before launch. Which solution meets these requirements?

1. A Assume the target is met if the design is multi-zone.
2. B Measure availability after launch.
3. C Model the availability of each component and its dependencies, then validate the model with fault injection testing.
4. D Estimate availability from the components' published figures.

142. A new workload must serve a read-heavy API with predictable sub-10-millisecond latency. Which solution meets these requirements?

1. A Place an in-memory cache in front of the data store and serve cached reads from it.
2. B Add read replicas to the data store.
3. C Increase the data store's provisioned IOPS.
4. D Increase the data store's instance size.

143. A new workload must process a large dataset in parallel with results combined at the end. Which solution meets these requirements?

1. A Increase the instance's memory allocation.
2. B Fan the work out across many workers with a coordinator that aggregates the results.
3. C Process the dataset sequentially on one large instance.
4. D Process the dataset on a schedule in small batches.

144. A new workload's static assets must be delivered with the lowest latency to a global audience. Which solution meets these requirements?

1. A Serve the assets from the origin Region.
2. B Replicate the assets to a bucket in every Region.
3. C Compress the assets before serving them.
4. D Distribute the assets through a content delivery network so they are cached near users.

145. An architect must select a data store for a workload requiring flexible queries across many attributes. Which solution meets these requirements?

1. A Choose a key-value store and scan for non-key attributes.
2. B Choose a store supporting secondary indexes or a query engine over the data, matched to the access patterns.
3. C Choose a store based on its published throughput figure.
4. D Choose the store the team has used previously.

146. A new workload's cost must be estimated before any resources are provisioned. Which solution meets these requirements?

1. A Provision the workload and observe the first month.
2. B Estimate from a similar workload's bill.
3. C Model the expected usage of each component and price it with the AWS Pricing Calculator, including data transfer.
4. D Assume cost scales linearly with user count.

147. A new workload must avoid paying for capacity during a lengthy development period. Which solution meets these requirements?

1. A Purchase Reserved Instances for the development environment.
2. B Use the largest instance types to speed development.
3. C Provision the production-sized environment from the start.
4. D Use serverless or on-demand services during development so cost follows actual usage.

148. An architect must decide whether a managed service justifies its price over a self-managed equivalent. Which solution meets these requirements?

1. A Choose the self-managed option because it appears cheaper.
2. B Compare only the hourly service charges.
3. C Compare the total cost including the staff effort, risk, and opportunity cost of operating the self-managed option.
4. D Choose the managed service because it is easier.

149. A new workload's storage cost must be minimized without knowing the access pattern in advance. Which solution meets these requirements?

1. A Use S3 Glacier Deep Archive for everything.
2. B Use S3 Intelligent-Tiering so objects move between tiers based on observed access.
3. C Use S3 Standard for everything.
4. D Apply a lifecycle policy with fixed transitions.

150. An architect must build cost awareness into a new workload's design. Which solution meets these requirements?

1. A Tag resources for attribution from the outset and define the cost metrics the team will track.
2. B Review the bill after the first month.
3. C Set a budget alert at the account level.
4. D Assign cost responsibility to the finance team.