AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 93
Question 920
Two product teams share a central platform account. Finance has categorized the platform support and data-transfer charges as SharedPlatform but wants an internal showback view allocating that shared category 70% to Team A and 30% to Team B. The real AWS bill and underlying CUR records must not be rewritten. Which AWS Billing capability matches?
- Activate an EC2 resource tag on Team A instances and assume it automatically splits platform-level support charges in the requested fixed percentages.
- Change the payer account’s payment instrument so AWS issues separate provider invoices for the 70% and 30% portions.
- Create two AWS Budgets with different alarm thresholds and expect those thresholds to reallocate SharedPlatform historical billing line items.
- Define an AWS Cost Categories split-charge rule using SharedPlatform as the source, both team categories as targets, and a fixed 70/30 allocation; inspect allocated figures in the cost-category details view.
Explanation
Correct Answer
D:Cost Categories supports fixed-percentage split-charge rules for shared source values. The allocated display is available on the cost-category details page; AWS notes that split results do not change CUR, Cost Explorer or the underlying bill.
Incorrect Answers
A:Ordinary resource cost-allocation tags help attribute eligible resource charges but do not automatically distribute a shared support category between two recipients at an arbitrary fixed ratio.
B:Changing the payer payment method does not construct a business-unit allocation report or change who incurred the shared platform line items in AWS billing data.
C:Budgets monitor or alert on spending, but their thresholds do not split a shared cost-category value or modify AWS Cost and Usage Report line items.
Question 921
A DynamoDB table has a very stable 5,000-read/1,000-write-per-second pattern all day. The team uses on-demand billing and can accurately forecast the load, with small periodic variance. Which billing alternative should it price?
- Evaluate provisioned read/write capacity with suitable auto scaling against observed steady request volume.
- Assume on-demand is invariably cheaper for every stable high-volume table.
- Set provisioned capacity far below the sustained load with no autoscaling or throttling plan.
- Move the table to S3 Glacier Deep Archive to retain DynamoDB's millisecond keyed-write semantics.
Explanation
Correct Answer
A:A well-utilized provisioned baseline can be less expensive than request-based capacity for predictable demand.
Incorrect Answers
B:Mode economics depend on utilization and price dimensions.
C:Underprovisioning can throttle business requests and destroy the anticipated benefit.
D:An archival object tier cannot replace operational DynamoDB API behavior.
Question 922
A DynamoDB provisioned table reserves its Black Friday read peak every day even though traffic is much lower most weeks. Billing is high and metrics show long idle periods. What configuration can improve capacity utilization?
- Keep the provisioned maximum permanently even when CloudWatch shows no requests.
- Configure provisioned autoscaling targets and sensible minimum/maximum RCU/WCU bounds from measured usage.
- Switch to a smaller partition key only to reduce cost without checking hot-key skew.
- Disable request metrics so reserved capacity appears fully used.
Explanation
Correct Answer
B:Provisioned capacity that adjusts with demand can avoid paying for peak units all year while retaining headroom.
Incorrect Answers
A:Provisioned database throughput or capacity can incur charges even when traffic is quiet, depending on the service and billing mode. Holding the maximum allocation through idle periods prevents cost reduction and should be evaluated against actual demand.
C:Key design changes can worsen throttling and do not automatically alter provisioned charges.
D:Hiding measurement does not lower the provisioned bill.
Question 923
A DynamoDB table is throttling reads to one extremely popular celebrity account despite table-wide spare capacity. Finance proposes purchasing huge additional RCUs for every partition. Which analysis should come first?
- Assume available aggregate RCUs can always be consumed entirely by one fixed partition key.
- Increase the size of each item to spread reads automatically across DynamoDB partitions.
- Investigate hot-key distribution and cache or key-splitting patterns before buying excess aggregate capacity.
- Add more global secondary indexes unrelated to the hot access path.
Explanation
Correct Answer
C:Single hot partitions can limit request throughput independently of idle table-wide units, making indiscriminate purchases inefficient.
Incorrect Answers
A:Partition-level throughput constraints can still produce hot-key throttling.
B:Larger items tend to consume more read capacity per operation.
D:Extra indexes cost storage/write capacity and may not resolve the hot key.
Question 924
A DynamoDB table stores temporary authentication challenges with TTL set to expire after ten minutes. Security policy insists the item must be provably erased at exactly 600 seconds. What limitation needs attention?
- Assume TTL immediately removes each item at the timestamp with millisecond guaranteed precision.
- Set a GSI sort key to the expiry timestamp and assume this enforces immediate physical deletion.
- Increase DynamoDB on-demand capacity and assume it changes the service's TTL timing guarantee.
- DynamoDB TTL deletion is asynchronous, so enforce expiry in application authorization and use stronger deletion controls where exact timing matters.
Explanation
Correct Answer
D:TTL marks items for eventual removal; an expired item can still exist briefly and should not be trusted as an active credential.
Incorrect Answers
A:TTL is an eventual background deletion process, not an exact-time secure erase service.
B:Index ordering does not make TTL erasure synchronous.
C:Read/write billing mode does not turn eventual expiry into deterministic destruction.
Question 925
A product service permanently stores every regenerable price preview in a large DynamoDB table, even though previews can be rebuilt from the current price source and expire within hours. Which data-retention change targets waste?
- Set suitable item TTL or application cleanup for expired reconstructible previews, while retaining the authoritative price source.
- Keep all previews forever because DynamoDB storage charges never increase with retained item count.
- Delete the authoritative price record and retain only expirable previews.
- Add multiple GSIs for the same transient fields solely to remove storage cost.
Explanation
Correct Answer
A:Removing unneeded cached data controls storage growth without deleting core business records.
Incorrect Answers
B:Billed stored bytes can grow with unnecessary historical entries.
C:That reverses the source-of-truth relationship and threatens correctness.
D:Indexes usually add storage and write overhead rather than removing expired base data.
Question 926
A DynamoDB table has seven GSIs. Every item write updates several projected attributes, causing more write capacity and storage charges than expected. Which cost component should be reviewed?
- Assume all GSIs are free views over the base table and never affect storage or write billing.
- Audit each GSI's necessity, projected attributes and write/read patterns before keeping all indexes.
- Increase the count of unused GSIs to guarantee fewer billed writes.
- Delete the table's partition key as a way to avoid GSI costs without redesigning access.
Explanation
Correct Answer
B:GSIs store separate indexed data and can consume additional write capacity when base items change.
Incorrect Answers
A:Indexes have their own storage and access capacity dimensions.
C:Each additional DynamoDB global secondary index consumes storage and index maintenance writes where applicable. Creating unused indexes can increase write units and cost rather than lowering the price of the original workload.
D:DynamoDB tables require valid primary key design and cannot simply abandon it.
Question 927
A retailer runs a DynamoDB Standard table in provisioned capacity mode in one AWS Region. Its read and write utilization has been stable for 18 months, and finance expects similar throughput for the next year. The team wants a discounted capacity commitment without changing the application data model. Which cost mechanism is eligible for this workload?
- Purchase reserved capacity for on-demand request units and apply the discount to any Standard-IA table in the organization.
- Buy DynamoDB reserved capacity in an inexpensive Region and let it automatically discount equivalent throughput in every other Region and AWS account.
- Evaluate a DynamoDB Reserved Capacity commitment for eligible provisioned RCUs and WCUs on the Standard table in its own Region, using the measured steady baseline.
- Reserve replicated write capacity units for all global-table replicas and apply that commitment to the table regardless of capacity mode.
Explanation
Correct Answer
C:DynamoDB Reserved Capacity offers discounted eligible provisioned read/write capacity units on DynamoDB Standard tables in the purchase Region, subject to term and minimums. Stable measured use supports evaluating the financial commitment.
Incorrect Answers
A:DynamoDB Reserved Capacity does not apply to on-demand mode or Standard-IA tables. It discounts eligible provisioned throughput for Standard tables and requires a committed use level.
B:Reserved Capacity is Region-scoped and not a portable multi-Region or transferable cross-account discount. The commitment should be matched to the actual eligible Regional throughput.
D:Replicated write capacity units are not eligible for DynamoDB Reserved Capacity, and the discount does not follow a table into a non-provisioned mode.
Question 928
A team's DynamoDB cost model includes base-table storage only. The workload adds a large GSI and uses change streams to feed another service. Which components should enter the revised estimate?
- Price the base-table WCU/RCU and storage but omit all index write amplification and its extra storage.
- Include GSI storage alone and assume every index Query uses free base-table read capacity.
- Assume change-stream event consumers never produce any billed reads, Lambda work or downstream service charges.
- Include index storage and write/read capacity, plus relevant DynamoDB Streams or change-data capture processing charges.
Explanation
Correct Answer
D:Cost the GSI projected attribute storage, index write/read activity and relevant stream/consumer data processing. A change stream and secondary index add usage dimensions beyond the base table even if the logical entity is unchanged.
Incorrect Answers
A:A GSI stores indexed entries separately and can consume additional write capacity and storage for inserts and updates.
B:A GSI has its own billed read capacity or on-demand request usage; paying only for bytes misses index query volume.
C:Streams and consuming integrations have service-specific API, processing and delivery costs; they cannot be excluded without checking the chosen design.
Question 929
A SaaS platform uses DynamoDB on-demand mode for highly variable tenant traffic. The platform must retain pay-per-request operation, but the team wants to limit exposure if a defective client suddenly issues far more reads and writes than the application can afford. Controlled throttling above an agreed request rate is acceptable. Which DynamoDB configuration best addresses this requirement?
- Set maximum on-demand read and write throughput values on the relevant table and GSIs, monitor throttling and costs, and treat the settings as best-effort throughput targets rather than absolute billing caps.
- Create an AWS Budgets alert and assume DynamoDB automatically rejects all requests the instant forecast charges exceed the alert threshold.
- Increase the table’s on-demand warm throughput and expect higher available throughput to impose a lower cost limit on runaway clients.
- Enable DynamoDB TTL on all items and assume their eventual expiration will stop excess reads and writes immediately during a request storm.
Explanation
Correct Answer
A:DynamoDB supports configurable maximum on-demand read and write throughput on tables and GSIs. Requests above these settings can be throttled, creating a practical cost-control guardrail; AWS notes burst capacity can temporarily exceed a configured maximum, so it is not a guaranteed hard currency cap.
Incorrect Answers
B:AWS Budgets alerts notify about costs or forecasts but are not themselves a synchronous DynamoDB request-throttling enforcement point.
C:Warm throughput describes immediately sustainable read/write capability. Increasing it does not create a lower request limit or constrain a defective client’s consumption.
D:TTL removes eligible records asynchronously after their expiration timestamp. It does not rate-limit incoming application requests or guarantee immediate deletion during a burst.
Question 930
An RDS MySQL db.m-family instance runs 24/7 but averages 4% CPU and modest memory use. Performance is normal even during peak weeks. What cost improvement is reasonable?
- Move the database to a burstable instance based only on four-percent average CPU, without examining baseline credits or peak-period memory demand.
- Profile peak CPU, memory, I/O, storage and connection behavior; load-test a smaller supported RDS class before committing to a right-size change.
- Reduce only the RDS allocated storage size in place and leave compute unchanged, because storage downsizing is the standard way to address low instance CPU.
- Replace Multi-AZ with a Single-AZ database as the first step without checking availability requirements or whether standby capacity causes the compute waste.
Explanation
Correct Answer
B:Right-sizing requires verifying full workload limits and a realistic peak, not just a CPU average. A tested smaller instance can reduce ongoing DB instance-hour cost while retaining target SLAs.
Incorrect Answers
A:Average CPU alone cannot establish that a burstable class has adequate baseline, credit and peak performance for a 24/7 database. This might replace an oversized host with a throttled one.
C:RDS allocated storage generally cannot be reduced in place, and storage size is a separate dimension from the overprovisioned CPU/memory of the DB instance class.
D:Availability topology should reflect explicit RTO/RPO and zone-failure requirements. Removing Multi-AZ solely to cut cost may violate resilience while not addressing instance sizing correctly.
