Amazon AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 76

Question 750

A table uses on-demand pricing for uniformly high, stable throughput and has a larger request bill than a provisioned forecast. Why?

Per-request on-demand billing can cost more than efficiently utilized provisioned capacity for steady traffic.
DynamoDB on-demand always costs less than provisioned capacity when a table has consistent high request volume.
DynamoDB charges exactly the same on-demand request-unit rate as the hourly provisioned throughput rate, regardless of utilization.
Adding auto scaling to provisioned mode guarantees a larger bill than on-demand for every steady read/write workload.

Explanation

Correct Answer

Calculate effective utilization and auto scaling headroom to choose the lower-cost mode.

Incorrect Answers

On-demand bills per request, and stable high throughput can justify efficiently utilized provisioned capacity.

On-demand request pricing and provisioned RCU/WCU pricing are different models. High, predictable demand can favor effectively utilized provisioned capacity.

Provisioned plus auto scaling can remain economical for stable demand if the chosen baseline and peak utilization justify it. Compare real workload rates.

Question 751

An API team switches a high-read DynamoDB table to Standard-IA and sees total costs rise. Why?

Standard-IA automatically reduces read/write request cost for a hot DynamoDB table because the storage price is lower.
Move the high-traffic DynamoDB workload to Standard-IA and add extra read-heavy GSIs while estimating savings using only the lower per-GB table storage rate.
Standard-IA request charges can outweigh storage savings when read/write volume is high.
Standard-IA automatically compresses and eliminates every GSI write and read charge when the base table becomes active.

Explanation

Correct Answer

Choose table class by storage/request cost mix rather than label alone.

Incorrect Answers

The infrequent-access table class can have higher per-request prices; frequent activity may outweigh storage savings.

Indexes and frequent requests can dominate spend. Lower Standard-IA storage cost is not enough to justify ignoring per-request throughput and index costs for active tables.

GSI activity still consumes billable resources; Standard-IA is a storage/request-pricing tradeoff, not a waiver for indexed operations.

Question 752

A service writes TTL values 24 hours in the past but still sees some expired items for a short time. Why?

DynamoDB TTL synchronously removes every expired item at precisely its timestamp and guarantees no stale item can be read one second later.
Writing an expiration timestamp alone performs a synchronous DynamoDB delete at the exact second, even if TTL is disabled on the table.
The TTL background process guarantees expired items cannot appear in Query responses while they await physical deletion.
DynamoDB TTL deletion is asynchronous/best effort; expired items may remain until background cleanup runs.

Explanation

Correct Answer

Filter expired logical records on reads when immediate invisibility is a business requirement.

Incorrect Answers

TTL deletion is asynchronous and eventual; expired items can remain temporarily visible until background cleanup.

DynamoDB must have TTL configured on the specified numeric epoch attribute, and its background deletion is asynchronous rather than instantaneous.

TTL-eligible items can remain visible until removal. Applications requiring immediate invisibility should filter or reject expired timestamps on reads.

Question 753

An RDS Reserved Instance purchase gives no benefit after the company changes to an ineligible engine/class/Region combination. Why?

The new database usage does not satisfy the reservation’s matching attributes or available flexibility rules.
An RDS Reserved Instance applies its benefit to every arbitrary database engine and Region after a one-time purchase.
Purchase an additional RDS Reserved DB instance in the original Region and expect the older reservation to become transferable to the new Region after the second purchase.
The reservation automatically converts itself to the new ineligible Region and DB class with no modification or remaining commitment.

Explanation

Correct Answer

Evaluate engine/class and regional eligibility before modifying committed database deployment.

Incorrect Answers

Reserved DB discounts require matching eligible attributes and flexibility rules; a move to an ineligible configuration can lose coverage.

Buying another reservation does not change which engine, class, Region and flexibility rules apply to the first reservation. Confirm matching attributes before treating a reservation as useful.

Reserved DB instance billing does not universally follow arbitrary migrations. Review eligible flexibility and replacement strategy before changing production usage.

Question 754

An Aurora Serverless v2 cluster maintains a high minimum ACU year-round despite almost no queries overnight. Why is the bill high?

Aurora Serverless v2 guarantees a zero-cost paused state for every compatible database engine and every minimum ACU setting.
Set the minimum ACUs to a nonzero value chosen for the daytime workload and assume auto-pause ignores that floor because the maximum ACU is large.
A high configured minimum capacity creates a billed compute floor even in otherwise idle periods.
Raising the Aurora Serverless maximum ACU forces the minimum to zero and eliminates every idle database compute charge.

Explanation

Correct Answer

Assess supported scale-to-zero options and minimum ACUs against resume-latency requirements.

Incorrect Answers

Auto-pause depends on supported versions and configuration; an elevated minimum ACU can retain a billed capacity floor.

A nonzero supported minimum can keep serverless compute running at a billable floor; maximum ACU defines a scale-up ceiling and cannot override minimum/auto-pause eligibility.

Maximum capacity controls the upper scaling limit. It does not override a high configured minimum or guarantee eligible auto-pause behavior.

Question 755

An RDS test instance that was stopped last week is running again although no operator explicitly started it. Why?

RDS stop/start always deletes the database storage and all snapshots, eliminating every recurring charge while preserving data.
Every stopped RDS DB instance remains stopped indefinitely until a human manually starts it, with no service-enforced limit.
RDS uses an automatic restart only if the database has a read replica, so a standalone test instance cannot resume by itself.
RDS automatically restarts certain stopped DB instances after the maximum supported continuous stop duration, commonly seven days.

Explanation

Correct Answer

Use lifecycle automation to stop it again when supported rather than assuming stopped forever.

Incorrect Answers

Stopping eligible RDS compute does not erase retained storage or backups, which may continue to accrue charges.

Amazon RDS can automatically restart an eligible stopped DB instance after the maximum consecutive stop interval, commonly seven days.

Supported stop/start limits apply independently of read replicas; absence of replicas does not guarantee indefinite stop.

Question 756

After creating three RDS read replicas, the monthly DB bill jumps but application queries still go to the primary. Why did replicas not deliver value?

The replicas incur ongoing charges while application connections were never moved to them.
Adding several RDS read replicas automatically redirects all application queries away from the primary without changing any connection strings.
Enable Multi-AZ standby for the primary and expect that synchronous availability standby to accept all dashboard read queries in place of the separate read replicas.
RDS read replicas incur no ongoing compute bill until application clients connect to their endpoints.

Explanation

Correct Answer

Verify read-routing demand and avoid spare replicas without an actual service objective.

Incorrect Answers

Replicas consume resources and need explicit client routing; unchanged application endpoints can leave paid replicas idle.

For a traditional RDS Multi-AZ DB instance, the standby is not an ordinary read endpoint. Read-scaling still requires supported read replicas and application routing to them.

Replica resources are provisioned and billed while running even if unused; check routing and workload need before adding replicas.

Question 757

A low-I/O Aurora cluster switches to I/O-Optimized and total charges rise. What explains the result?

Aurora I/O-Optimized always has a lower total bill than Aurora Standard at any read/write I/O level.
When I/O usage is modest, the different baseline compute/storage pricing can cost more than Aurora Standard with low I/O fees.
Aurora I/O-Optimized always costs less than Standard when database read/write I/O usage is nearly zero.
Enabling I/O-Optimized forces Aurora to discount all allocated cluster compute and storage by the same percentage.

Explanation

Correct Answer

Choose based on measured I/O-to-total cost ratio rather than always turning on I/O-Optimized.

Incorrect Answers

Different baseline compute/storage and I/O charges mean the crossover depends on the real workload volume.

I/O-Optimized trades different compute/storage rates for no separately metered read/write I/O. Low I/O can favor Standard.

I/O-Optimized is a pricing configuration with different component rates; it is not an unconditional discount on all resources.

Question 758

A cache cluster is provisioned but the database cost does not fall because every update invalidates the entire cache. Why?

An ElastiCache cluster automatically intercepts all RDS SQL queries even when the application always queries the relational database first.
Broad invalidation causes persistent cache misses and leaves most reads reaching the database.
Invalidating every cache key after any update increases cache hits for unchanged product records.
The database automatically stops charging for SQL reads merely because an ElastiCache cluster exists, even if queries bypass it.

Explanation

Correct Answer

Use granular invalidation/TTL with consistency guarantees and check hit-rate metrics.

Incorrect Answers

Cache-aside is application integration; if every request hits RDS first, a provisioned cache offers little relief.

Broad invalidation discards reusable data and raises miss rates. Invalidate affected keys selectively and verify measured hit ratio.

The application must actually retrieve repeated data from a warm cache; misses and bypassed reads continue loading the database.

Question 759

A mobile service added two DynamoDB global-table replica Regions to improve local reads. The replicas add storage and replicated-write charges, but telemetry shows every mobile client still reads from the original Region. What cost-benefit decision should the architect make?

Keep the added replicas automatically because their presence alone makes mobile clients resolve DynamoDB reads to the nearest Region.
Remove every global-table replica immediately because DynamoDB never supports Region-local application reads without cross-Region data transfer.
Reduce the original Region storage allocation by two-thirds and assume global tables automatically share one physical copy across all Regions.
Measure real Region-local read latency and routing needs, then retain only replicas whose local-read and recovery benefits justify replicated writes and regional storage.

Explanation

Correct Answer

The original goal was lower client latency, not mere extra copies. Verify that traffic is routed to and benefits from those Regions, then compare that measured value against regional replicated-write and storage costs.

Incorrect Answers

Global tables do not automatically redirect a client SDK or application to the closest replica. The application must select its intended regional endpoint and access path.

Global tables can serve requests from replicas in supported Regions. Whether a particular deployment requires them depends on its read routing, availability and consistency needs.

Every global-table replica maintains billed data and related indexes in its Region; adding replicas does not automatically split the source data into cheaper shared storage.

Question 760

An RDS storage autoscaling database reaches its configured maximum allocated storage and fails to grow further. Why?

RDS storage autoscaling can always exceed its configured maximum allocated storage without any intervention.
Storage autoscaling cannot increase allocated size beyond the configured maximum storage threshold.
Increase the autoscaling minimum allocated storage but leave the maximum allocated-storage ceiling equal to the present size, assuming the ceiling moves automatically.
Turning on an increased CloudWatch monitoring period raises the database storage auto scaling maximum without a DB configuration change.

Explanation

Correct Answer

Inspect free storage metrics and ceiling settings; autoscaling is not unlimited.

Incorrect Answers

Autoscaling is bounded by maximum storage settings and supported scaling requirements; reaching the cap blocks further automatic allocation.

The maximum allocation is an independent explicit limit. Raising a minimum or baseline does not guarantee the database will exceed the configured ceiling when storage pressure grows.

Metrics can reveal the pressure but cannot change the maximum allocated-storage setting; revise the allowed maximum after validating limits.

Pass Your Amazon AWS Certified Solutions Architect - Associate SAA-C03 Exam Easy!

Question 750

Explanation

Correct Answer

Incorrect Answers

Question 751

Explanation

Correct Answer

Incorrect Answers

Question 752

Explanation

Correct Answer

Incorrect Answers

Question 753

Explanation

Correct Answer

Incorrect Answers

Question 754

Explanation

Correct Answer

Incorrect Answers

Question 755

Explanation

Correct Answer

Incorrect Answers

Question 756

Explanation

Correct Answer

Incorrect Answers

Question 757

Explanation

Correct Answer

Incorrect Answers

Question 758

Explanation

Correct Answer

Incorrect Answers

Question 759

Explanation

Correct Answer

Incorrect Answers

Question 760

Explanation

Correct Answer

Incorrect Answers

AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests