Question 550
An organization runs Kubernetes nonproduction environments in EKS that are idle every night. The worker node groups continue running because minimum capacity is never reduced. Governance approves stopping workloads outside office hours. What cost improvement targets the waste?
- Schedule safe scale-down of nonproduction workloads and compatible node capacity, then restore baseline before business-hour use.
- Increase worker minimum capacity at night so empty clusters always have double the compute allocation.
- Scale only application Pod replicas to zero overnight while leaving idle EKS EC2 worker node groups at unchanged minimum capacity.
- Delete all EKS clusters nightly without preserving cluster state, network configuration or scheduled morning restoration.
Explanation
Correct Answer
A:Automated schedules reduce unused worker compute-hours while allowing planned exceptions and testing stateful constraints.
Incorrect Answers
B:Additional idle node capacity increases the specific measured nightly waste.
C:Scaling Pods down without adjusting node group capacity can leave the core idle EC2 compute charge intact.
D:Destructive recreation introduces availability and configuration risks rather than controlled nonproduction workload and node scheduling.
Question 551
AWS Compute Optimizer identifies EC2 instances whose measured CPU, memory and network usage are far below allocated sizes. Business owners require a no-surprises change with a rollback option. What should the platform recommend?
- Select a larger instance size than any AWS Compute Optimizer recommendation without validating peak workload memory or network requirements.
- Apply the smallest instance family across all accounts regardless of incompatible CPU architecture, network throughput, or peak memory.
- Validate recommendations using workload telemetry, test smaller instance types and deploy right-sizing changes with rollback.
- Apply every smallest available instance type directly in production without compatibility or peak-load tests.
Explanation
Correct Answer
C:Workload-informed sizing can remove sustained overprovisioning while preserving application limits and safe recovery.
Incorrect Answers
A:Larger instances ignore the evidence of waste and cannot be assumed to meet cost or reliability targets efficiently.
B:Blind fleet-wide downsizing can break workloads; recommendation validation must include application compatibility and rollback.
D:Blind size reductions risk OOM or throughput failures and ignore observed application requirements.
Question 552
An Amazon OpenSearch Service domain stores 18 months of search indices. Investigators query recent writes hourly, but indices older than 60 days are immutable and searched only during quarterly audits. The domain uses a supported engine and has compatible dedicated master nodes, and finance wants lower storage cost without making old indices completely unqueryable. Which storage tier should be evaluated?
- Enable OpenSearch Reserved Instances for the same expensive hot nodes and assume the storage tier changes automatically to UltraWarm with the reservation.
- Move eligible infrequently queried read-only indices from hot nodes to Amazon OpenSearch Service UltraWarm storage, validating supported performance and lifecycle constraints.
- Increase the hot-data EBS allocation and node size for all 18 months of infrequently accessed indices, expecting storage cost to decrease.
- Move active writable hot indices directly to S3 Glacier Deep Archive and assume OpenSearch supports regular near-real-time search over those archived objects.
Explanation
Correct Answer
B:UltraWarm uses cost-effective storage with a queryable cache for older immutable indices whose access rate does not justify hot-tier costs.
Incorrect Answers
A:Reserved compute discounts do not transfer historical immutable indices into a lower-priced warm storage tier.
C:Additional hot storage and compute add cost without using the lower-cost eligible searchable warm tier.
D:Glacier archival is not an ordinary online OpenSearch searchable index tier and cannot handle live writes or low-latency queries as stated.
Question 553
An Aurora MySQL workload performs many small database I/O operations. Its monthly bill is dominated by I/O request charges; CPU and storage are steady. What pricing configuration deserves a measured comparison before migration?
- Change the Aurora cluster instance class to a larger class but preserve Standard I/O pricing and the same high number of storage I/O requests.
- Use RDS Reserved DB Instances to discount eligible Aurora compute and assume the reservation changes the I/O request pricing model itself.
- Enable query logging in CloudWatch Logs and assume that action waives the Aurora database storage I/O billing line.
- Compare Aurora I/O-Optimized and Standard clusters against observed I/O charge volume and total monthly database costs.
Explanation
Correct Answer
D:Aurora I/O-Optimized can offer different I/O billing economics; comparison must account for compute, storage and actual I/O demand.
Incorrect Answers
A:Instance sizing may change compute costs without addressing whether I/O-Optimized has better economics for the observed dominant request charges.
B:Reservations can affect eligible compute economics but are not the storage configuration that switches Aurora between Standard and I/O-Optimized.
C:Additional monitoring is useful but does not change the chosen Aurora storage I/O pricing model.
Question 554
An ecommerce platform retains an 11-TiB DynamoDB OrdersHistory table for audit lookups. It uses DynamoDB Standard table class, but monthly billing shows persistent data and index storage dominates costs while reads and writes are rare; the application still needs ordinary key queries, the current durability level and low-latency responses. Finance wants to reduce storage expense without exporting the data to a retrieval-delay archive. What should be evaluated?
- Compare the DynamoDB Standard-Infrequent Access (Standard-IA) table class to Standard using measured table and index storage charges and its higher read/write request prices.
- Move the DynamoDB table directly into S3 Glacier Deep Archive and assume DynamoDB Query requests will still run against that archive with normal latency.
- Increase provisioned read and write units for the rarely accessed history table because larger reserved capacity automatically lowers the per-GB DynamoDB storage rate.
- Enable DynamoDB time to live on the required seven-year order records so the historical audit data is deleted automatically after an arbitrary seven-day period.
Explanation
Correct Answer
A:DynamoDB Standard-IA lowers storage cost for storage-dominated, infrequently accessed tables while retaining key-value service semantics; request charges differ, so total cost including indexes and actual throughput must be compared.
Incorrect Answers
B:S3 cold archive is a different object-storage retrieval workflow and does not preserve live DynamoDB queries or table behavior.
C:Extra throughput capacity adds cost and does not change the table class storage pricing of historical records.
D:Automatic early deletion of legally required records violates the retention requirement and does not preserve queryable history.
Question 555
An S3 bucket receives millions of small short-lived temporary exports. Most remain for months because teams never clean them up, and product owners confirm files may be deleted seven days after creation. Which policy reduces the recurring expense?
- Migrate all objects to S3 Glacier Deep Archive and assume early deletion has no minimum-duration charge implications.
- Enable S3 Inventory reports but leave the seven-day temporary exports indefinitely in the bucket without a deletion lifecycle.
- Configure an S3 lifecycle expiration rule targeting the approved temporary-export prefix after the seven-day retention window.
- Enable versioning without an expiration policy and assume all temporary versions immediately stop costing money.
Explanation
Correct Answer
C:Automated expiration safely removes objects once their approved useful lifetime ends and prevents neglected accumulation.
Incorrect Answers
A:Deep archive transitions can introduce minimum storage duration and retrieval costs for very short-lived exports.
B:Inventory helps locate stale objects but does not delete them or reduce storage charges without an applied expiration policy.
D:Versioning often retains additional object data unless lifecycle cleanup rules manage those copies.
Question 556
An established multi-AZ Amazon EFS file system stores 40 TiB of project files. Metrics show that most files older than 90 days are read only a few times per quarter, while recent files remain latency-sensitive. Data must stay in the existing managed shared NFS file system and occasional retrieval charges are acceptable. Which storage-cost optimization should be tested?
- Change the Amazon EFS throughput mode from Bursting to Elastic and assume that action automatically transitions historical file bytes into IA storage.
- Configure an EFS replication destination in another Region and assume replicating Standard-class bytes to a second file system lowers total storage charges.
- Re-create the EFS file system as One Zone to reduce rates while discarding the required multi-AZ durability and availability protections.
- Enable supported Amazon EFS Lifecycle Management policies to transition eligible infrequently accessed files to EFS IA or Archive, and compare storage savings with access charges.
Explanation
Correct Answer
D:EFS Lifecycle Management tracks file access and can transition less-active files into IA or Archive classes while retaining the managed shared-file interface, subject to supported eligibility and read charges.
Incorrect Answers
A:EFS throughput mode controls throughput behavior; it is independent of the lifecycle policies that place file data in Standard, Infrequent Access, or Archive classes.
B:EFS replication provides data availability rather than low-cost lifecycle tiering, and storing a second copy does not reduce the amount of Standard data in the source.
C:EFS One Zone changes the durability and availability domain; it is not an in-place lifecycle transition compatible with a requirement to preserve regional shared availability.
Question 557
A data-engineering team runs three compatible AWS Glue Spark ETL jobs every weekend to transform historical datasets. The jobs are not customer-facing, may start later and complete at variable times, and can be retried if spare compute is interrupted. Standard execution is the main driver of the ETL bill. The jobs use a supported Glue version and worker type and run in a Region offering Flex. Which execution-class choice should be tested?
- Purchase EC2 Reserved Instances for the Glue service and assume a separate EC2 capacity reservation automatically changes managed Glue Spark billing.
- Change the jobs to Glue streaming mode with nonstop execution, even though each job is a finite and nonurgent historical batch.
- Increase each Spark job worker count to the service maximum regardless of measured useful parallelism and startup delay.
- Use the AWS Glue FLEX execution class for eligible Spark ETL job runs, accepting variable start/completion times and possible spare-capacity interruptions.
Explanation
Correct Answer
D:Glue Flex is intended for eligible nonurgent Spark jobs that tolerate scheduling variability and spare-capacity interruptions, in exchange for lower execution cost compared with Standard.
Incorrect Answers
A:EC2 Reserved Instances do not directly turn a managed Glue job into a reserved-EC2 billed resource or choose the supported Glue execution class.
B:Continuous streaming execution is a different processing model and can increase running time for finite intermittent batch work.
C:Provisioning additional DPUs without utilization evidence can raise cost and does not exploit lower-cost flexible execution.
Question 558
Finance receives large monthly bills from idle Provisioned DynamoDB write capacity, but table write traffic has predictable long daily quiet periods. The business prefers retained provisioned billing over on-demand because peaks are stable. Which change is best to test?
- Use DynamoDB Application Auto Scaling with a measured target utilization and suitable min/max provisioned capacity.
- Keep the maximum write capacity provisioned around the clock regardless of the observed recurring quiet periods.
- Increase DynamoDB provisioned write capacity to the maximum for every predictable nightly quiet period, expecting higher idle WCU to lower charges.
- Disable all CloudWatch capacity telemetry and assume auto scaling can react accurately without a signal.
Explanation
Correct Answer
A:Capacity auto scaling can adjust provisioned units with sustained traffic patterns, lowering idle allocation while respecting safeguards.
Incorrect Answers
B:Fixed peak capacity pays for unused throughput during the predictable low-demand interval.
C:Greater provisioned WCU during idle hours moves cost in the wrong direction and does not implement utilization-driven capacity scaling.
D:Removing required utilization measurements undermines the managed scaling decision process.
Question 559
A cloud-finance team already receives unusual-spend alerts. It now wants one organization-wide view of estimated saving opportunities across EC2 rightsizing, idle resource cleanup, and discounted compute purchases, taking existing Savings Plans and Reserved Instances into account so recommendations are not double-counted. Which AWS cost-management capability best matches?
- Create another Cost Anomaly Detection alert for every account, because anomalous-spend notifications automatically reconcile all RI and rightsizing savings recommendations.
- Enable AWS Cost Optimization Hub for the organization and assess its aggregated prioritized savings recommendations using applicable commercial terms.
- Use AWS Compute Optimizer rightsizing recommendations alone without estimating unused commitments or savings overlaps across separate recommendations.
- Use AWS Budgets alert thresholds as the only program for combining and prioritizing rightsizing and commitment purchases across the organization.
Explanation
Correct Answer
B:Cost Optimization Hub consolidates and prioritizes eligible cost optimization recommendations across accounts and Regions and accounts for relevant RI/Savings Plans commercial terms.
Incorrect Answers
A:Anomaly alerts detect unexpected spend patterns but do not consolidate cost recommendation estimates across product types and commitments.
C:Compute Optimizer produces resource-level analysis; Cost Optimization Hub consolidates and deduplicates the broader discounted savings opportunities requested.
D:Budgets is a spending-alert or action facility, not an aggregation and overlap-aware savings recommendation engine.
Question 560
A large object archive already uses S3 Intelligent-Tiering and has automatically moved eligible files between Frequent, Infrequent and Archive Instant Access tiers. Millions of eligible objects remain untouched for 200 days or longer and are retained for years; the archive owner accepts asynchronous restore taking hours when a rare request arrives. Finance wants a deeper storage price reduction without migrating the objects out of their existing Intelligent-Tiering storage class. What tiering configuration should be considered?
- Set every object back to Frequent Access after 200 days and disable all optional archival configurations, even though almost none of those objects are retrieved.
- Enable supported Intelligent-Tiering optional Archive Access and/or Deep Archive Access tiers with appropriate 90/180-day inactivity thresholds, accounting for asynchronous restore and policy constraints.
- Assume the automatic Archive Instant Access tier has hour-long retrieval delays and is exactly the same as optional Deep Archive Access, so no changes are possible.
- Delete every object with no GET request for 200 days even though retention requirements require its data to remain recoverable.
Explanation
Correct Answer
B:Intelligent-Tiering can optionally archive eligible inactive data to Archive Access after a minimum 90 days or Deep Archive Access after a minimum 180 days, with restore delays that fit the specified infrequent access.
Incorrect Answers
A:Forcing dormant data into Frequent Access increases storage cost and does not use the eligible optional archival tiers.
C:Archive Instant Access provides millisecond access and is an automatic tier, whereas optional archival tiers require explicit configuration and asynchronous restore.
D:Unrequested objects may still be within the mandatory retention period; inactivity is not authority to delete them.
AWS Certified Solutions Architect - Professional SAP-C02 Practice Tests