Resilient Cloud Solutions

88. A stateful service must survive the loss of an Availability Zone with its data intact and no manual intervention. Which solution meets these requirements?

1. A Deploy the data layer with a read replica in a second zone for manual promotion.
2. B Deploy the compute layer across multiple zones with the data layer on instance storage.
3. C Deploy the data layer on a single instance with frequent snapshots to Amazon S3.
4. D Deploy the data layer as a Multi-AZ managed database with automatic failover, and the compute layer across multiple zones.

89. Single points of failure must be identified across many existing workloads rather than discovered during an incident. Which solution meets these requirements?

1. A Enable AWS Trusted Advisor and review its fault tolerance checks each quarter.
2. B Assess the workloads with AWS Resilience Hub and remediate the findings it reports against the defined resiliency policy.
3. C Wait for an incident and record the failure point in the post-incident review.
4. D Review each workload's architecture diagram during the annual design review.

90. An Amazon S3 bucket's contents must remain available if the primary Region becomes unavailable. Which solution meets these requirements?

1. A Configure S3 Cross-Region Replication to a bucket in a second Region and fail readers over to it.
2. B Create a lifecycle rule that transitions objects to an archival storage class.
3. C Enable S3 Intelligent-Tiering so objects move between access tiers automatically.
4. D Enable S3 Versioning on the primary bucket so previous versions are retained.

91. An Application Load Balancer must distribute traffic evenly across targets in three Availability Zones where target counts differ per zone. Which configuration meets these requirements?

1. A Enable cross-zone load balancing so requests are distributed across all registered targets regardless of zone.
2. B Enable sticky sessions so each client stays with one target.
3. C Reduce the load balancer's health check interval so unhealthy targets are removed sooner.
4. D Register an equal number of targets in each Availability Zone.

92. A business requirement states a service level of 99.95 percent availability for a workload currently deployed in one Availability Zone. Which step should be taken first?

1. A Purchase a support plan with a faster response time commitment.
2. B Identify the single points of failure in the current design and determine which must be removed to meet the target.
3. C Increase the frequency of backups so recovery is quicker.
4. D Move the workload to a larger instance type to reduce the chance of failure.

93. An Auto Scaling group scales on average CPU, but the workload is bound by queue depth rather than CPU. Which solution meets these requirements?

1. A Increase the group's maximum size so more capacity is available.
2. B Reduce the CPU target so the group scales out sooner.
3. C Scale on a custom metric such as backlog per instance derived from the queue depth.
4. D Shorten the scaling cooldown so scaling actions occur more often.

94. An Amazon ECS service must scale its tasks with demand while the underlying capacity scales to match. Which solution meets these requirements?

1. A Configure a fixed task count and scale the Auto Scaling group on instance CPU.
2. B Configure service auto scaling for the tasks and a fixed-size Auto Scaling group.
3. C Configure service auto scaling for the tasks and a capacity provider with managed scaling for the underlying capacity.
4. D Configure the service to run on AWS Fargate with a fixed task count.

95. A read-heavy workload's database is saturated by repeated identical queries. Which solution most reduces load on the database?

1. A Add read replicas and direct the repeated queries to them.
2. B Add a caching layer that serves the repeated query results without reaching the database.
3. C Increase the database's provisioned IOPS so storage responds faster.
4. D Increase the database instance class so queries complete faster.

96. A serverless application must handle a predictable morning traffic spike without cold start latency. Which solution meets these requirements?

1. A Configure provisioned concurrency on the function and schedule it to increase before the spike.
2. B Place an Amazon SQS queue in front of the function to buffer the spike.
3. C Configure reserved concurrency so the function is guaranteed capacity.
4. D Increase the function's memory allocation so initialization completes faster.

97. A workload must serve users on three continents with low latency and must tolerate the loss of a Region. Which solution meets these requirements?

1. A Deploy the workload in one Region with read replicas in the other two continents.
2. B Deploy the workload in one Region on larger instances to reduce processing time.
3. C Deploy the workload in multiple Regions and route users to the nearest healthy Region with latency records and health checks.
4. D Deploy the workload in one Region and serve static content through a content delivery network.

98. A workload must recover from Regional failure with a recovery time objective of fifteen minutes and a recovery point objective of five minutes. Which strategy meets these requirements?

1. A Implement an active-active deployment at full capacity in both Regions.
2. B Implement a pilot light with data replicated and the application tier provisioned at failover.
3. C Implement backup and restore from cross-Region snapshots.
4. D Implement a warm standby with continuous data replication and a scaled-down environment running in the second Region.

99. Failover testing must run on a schedule and produce evidence, rather than depending on an engineer initiating it. Which solution meets these requirements?

1. A Record the recovery procedure in the runbook and review it each quarter.
2. B Automate the failover exercise as a scheduled workflow that performs the failover, validates the result, and records the outcome.
3. C Confirm each night that the cross-Region replication lag is within the recovery point objective.
4. D Add a recurring calendar entry reminding the team to perform a failover exercise.

100. A backup policy must be applied uniformly across every account in an organization, including accounts created later, without per-account configuration. Which solution meets these requirements?

1. A Create an AWS Backup plan in each member account with the same settings.
2. B Deploy an AWS Backup plan to each account with CloudFormation StackSets and update the StackSet when the policy changes.
3. C Create a Lambda function that creates a backup plan whenever a new account is detected.
4. D Create an AWS Backup policy in AWS Organizations and attach it to the organizational unit so member accounts inherit it.

101. A stateful service must fail over between Availability Zones without losing in-flight data. Which approach is appropriate?

1. A Replicate the state synchronously to a standby in another zone so a failover loses no committed data.
2. B Snapshot the state every five minutes to another zone.
3. C Replicate the state asynchronously and accept the replication lag.
4. D Rebuild the state from source systems after a failover.

102. A cross-Region solution must keep an Amazon DynamoDB table readable and writable in two Regions. Which solution meets these requirements?

1. A Enable point-in-time recovery and restore into the second Region during failover.
2. B Copy the table to the second Region with a nightly export.
3. C Configure the table as a global table so both Regions accept writes and replicate to each other.
4. D Create a read replica of the table in the second Region.

103. A load balancer must continue serving when one Availability Zone's targets all become unhealthy. Which configuration matters most?

1. A Targets registered in more than one zone, so the load balancer has healthy destinations elsewhere.
2. B A shorter health check interval so failures are detected sooner.
3. C A longer deregistration delay so connections drain fully.
4. D Sticky sessions so clients remain with their assigned target.

104. An Auto Scaling group launches instances that take eight minutes to become ready, and scaling lags demand. Which improvement is appropriate?

1. A Lower the scaling threshold so scaling begins sooner.
2. B Increase the group's maximum size so more capacity is available.
3. C Reduce the scaling cooldown so scaling actions occur more often.
4. D Reduce instance startup time with a pre-baked image and use predictive or scheduled scaling ahead of known demand.

105. A container platform must scale its worker nodes only when pods cannot be scheduled. Which approach is appropriate?

1. A Maintain a fixed node count sized for peak pod demand.
2. B Scale the node group in response to unschedulable pods rather than to node resource utilization.
3. C Scale the node group on average CPU utilization across the nodes.
4. D Scale the node group on a fixed schedule matching business hours.

106. A serverless application's downstream database is overwhelmed when the function scales out. Which solution meets these requirements?

1. A Increase the function's timeout so slow database calls complete.
2. B Increase the function's memory allocation so invocations complete faster.
3. C Set reserved concurrency on the function to cap its scale, and pool database connections through a proxy.
4. D Increase the database instance size to absorb the connection load.

107. A workload must be deployed to several Regions with one definition and one deployment action. Which approach is appropriate?

1. A Define the workload as infrastructure as code and deploy it with StackSets targeting the Regions.
2. B Deploy in one Region and replicate the resources with a custom script.
3. C Copy the template and deploy it manually in each Region.
4. D Deploy in one Region and route global traffic to it.

108. A recovery procedure must be validated to meet a stated recovery time objective. Which approach is appropriate?

1. A Execute the procedure under realistic conditions and measure the elapsed time against the objective.
2. B Estimate the elapsed time from the documented steps.
3. C Confirm that backups complete within the objective's window.
4. D Confirm that the standby environment's resources exist.

109. An automated recovery must restore a database to a point immediately before a bad deployment. Which capability is required?

1. A Point-in-time recovery enabled on the database before the incident occurs.
2. B Automated snapshots taken daily.
3. C A read replica promoted at the time of the incident.
4. D Cross-Region backup copies of the most recent snapshot.

110. An Aurora cluster must fail over to a replica with minimal disruption when the writer instance fails. Which configuration is appropriate?

1. A Add a reader instance with a failover priority and connect applications through the cluster endpoint.
2. B Connect applications directly to the writer instance's endpoint.
3. C Take frequent snapshots of the cluster.
4. D Increase the writer instance's size.

111. A workload must survive the loss of an entire Region. Which design is required?

1. A Deploy the workload and replicate its data to a second Region, with a mechanism to direct traffic there.
2. B Take cross-Region snapshot copies without a standby environment.
3. C Increase the workload's instance count in the primary Region.
4. D Deploy the workload across three Availability Zones in one Region.

112. An Elastic Load Balancer must stop routing to a target whose dependencies have failed even though the target process is running. Which configuration is appropriate?

1. A Reduce the health check interval.
2. B Increase the unhealthy threshold.
3. C Point the health check at the target's TCP port.
4. D Point the health check at an endpoint that verifies the target's dependencies and returns unhealthy when they fail.

113. An Auto Scaling group must replace instances gradually when a new launch template version is published. Which capability is appropriate?

1. A Doubling the desired capacity and halving it afterwards.
2. B Manually terminating instances one at a time.
3. C Terminating all instances so the group relaunches them.
4. D An instance refresh with a minimum healthy percentage so instances are replaced in controlled batches.

114. A scaling policy must add capacity proportionally to how far a metric exceeds its threshold. Which policy type is appropriate?

1. A A scheduled scaling action.
2. B A simple scaling policy with a fixed adjustment.
3. C A target tracking policy on the metric.
4. D A step scaling policy with steps defined for increasing breach magnitudes.

115. A predictable weekly traffic pattern must be met without waiting for reactive scaling. Which capability is appropriate?

1. A Target tracking on CPU utilization.
2. B Step scaling with aggressive thresholds.
3. C A larger instance type.
4. D Predictive scaling, which forecasts from historical patterns and provisions capacity in advance.

116. A recovery process must restore an application and its data in a second Region with the least manual work. Which approach is appropriate?

1. A Keep a spreadsheet of the resource configuration.
2. B Define the environment as infrastructure as code deployed to the second Region, with data replicated continuously.
3. C Document the manual rebuild steps in a runbook.
4. D Take snapshots and rebuild the environment by hand during an incident.

117. An AWS Backup plan must copy recovery points to a second Region and retain them for a defined period. Which configuration is appropriate?

1. A Copy the recovery points manually each month.
2. B Enable cross-Region replication on the source storage.
3. C Add a copy action to the backup rule targeting a vault in the second Region with its own retention.
4. D Create a separate backup plan in the second Region.

118. An Aurora cluster's reader endpoint must distribute read traffic across replicas. Which configuration is required?

1. A Create a load balancer in front of the replicas.
2. B Connect read queries to the cluster's reader endpoint, which balances across available replicas.
3. C Connect read queries to the writer endpoint.
4. D Connect read queries to an individual replica's instance endpoint.

119. A multi-Region active-passive workload must detect that the primary Region is unhealthy from outside that Region. Which approach is appropriate?

1. A Run health checks from a service that evaluates the endpoint globally rather than from within the primary Region.
2. B Rely on the primary Region's CloudWatch alarms.
3. C Run health checks from an instance in the primary Region.
4. D Ask users to report outages.

120. A service must remain available when a dependency in another Availability Zone becomes slow rather than failing outright. Which pattern applies?

1. A A longer request timeout.
2. B An increased connection pool size.
3. C An unbounded retry until the dependency responds.
4. D A timeout with a fallback, so a slow dependency does not hold resources indefinitely.

121. An ElastiCache cluster must survive the loss of a node without losing cached data. Which configuration is appropriate?

1. A Use a single-node cluster with frequent snapshots.
2. B Increase the node's memory allocation.
3. C Distribute the cluster's nodes across Regions.
4. D Use a replication group with Multi-AZ enabled and automatic failover so a replica is promoted.

122. An Auto Scaling group must avoid scaling on a metric spike caused by instances starting up. Which configuration is appropriate?

1. A Raise the scaling threshold.
2. B Reduce the group's maximum size.
3. C Configure a warm-up period so new instances are excluded from the group's aggregate metric until they are ready.
4. D Increase the scaling cooldown period.

123. A serverless application's downstream API allows a fixed number of concurrent connections. Which configuration bounds the load?

1. A Increase the function's memory allocation.
2. B Set reserved concurrency on the function so its concurrent executions cannot exceed the downstream limit.
3. C Add a retry policy to the function.
4. D Increase the function's timeout.

124. An ECS service must scale on a metric published by the application rather than on infrastructure utilization. Which configuration is appropriate?

1. A Configure a fixed task count sized for peak.
2. B Configure target tracking on the service's average CPU utilization.
3. C Configure scheduled scaling at fixed times.
4. D Configure service auto scaling with a target tracking policy on the custom application metric.

125. A recovery process must be validated without an engineer initiating it each time. Which approach is appropriate?

1. A Document the recovery steps and review them annually.
2. B Perform the exercise after each production incident.
3. C Schedule the recovery exercise as an automated workflow that performs, validates, and records the outcome.
4. D Add the exercise to the team's quarterly calendar.

126. An automated recovery must not trigger during a planned maintenance window. Which approach is appropriate?

1. A Check for an active maintenance window in the recovery automation and exit without acting during one.
2. B Accept the recovery firing and let it complete.
3. C Extend the alarm's evaluation period to cover the window.
4. D Disable the alarms before each maintenance window and re-enable them afterwards.