Amazon AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 80

Question 790

An S3 replication rule uses Replication Time Control (RTC) for every log object, but only compliance records need the documented 15-minute replication SLA. Most verbose debug objects can replicate without that SLA. Which optimization retains the critical guarantee?

Turn off all S3 replication and rely on Route 53 weighted DNS answers to maintain a compliant secondary copy of the critical objects.
Enable S3 RTC on the noncritical debug prefix only and disable replication for compliance records, assuming the RTC price is lowered by reversing the scope.
Change the replication destination bucket storage class but leave RTC enabled for every debug object, assuming the per-object RTC feature charge is eliminated.
Use rule filters to apply RTC only to compliance-critical prefixes or tags; use standard replication for eligible other objects and validate the lower-priced SLA tradeoff.

Explanation

Correct Answer

S3 supports prefix/tag-filtered replication rules and RTC enabled on the selected rule. Limit RTC to data that needs the 15-minute replication objective and validate standard replication for other data, preserving all mandatory secondary copies.

Incorrect Answers

DNS routing determines which application endpoint a client resolves and does not replicate S3 objects. Removing required replication fails the documented compliance recovery objective.

RTC provides the shorter documented replication-time promise to the objects covered by its rule. Assigning it to noncritical objects while excluding critical records reverses the business requirement.

Storage class economics and RTC replication features are distinct. Merely changing destination storage class cannot eliminate the charges for RTC replication still enabled on noncritical objects.

Question 791

A team pays for a large Direct Connect port but uses only a few Mbps quarterly. The economics are poor. Why?

The provisioned Direct Connect connection has continuing port/partner costs that may outweigh savings at low utilization.
Direct Connect port-hour charges stop automatically on days that the provisioned private circuit carries no packets.
Enable Direct Connect virtual interface BGP keepalives more frequently and assume this causes the provisioned port’s hourly price to scale with transferred bytes.
Move more VPC route prefixes onto the same Direct Connect virtual interface and assume advertised route count reduces the fixed port subscription charge.

Explanation

Correct Answer

Compare actual duty cycle and provider charges rather than treating a dedicated line as universally cheaper.

Incorrect Answers

Provisioned dedicated connectivity can incur continuing port or partner costs even when the link is underused.

BGP keepalive timing does not convert a provisioned circuit-hour charge to a pure data-usage charge.

BGP route advertisement does not automatically lower port, hosted-connection or partner recurring charges.

Question 792

A company has migrated from an AWS Direct Connect dedicated circuit to VPN. It deletes the Direct Connect virtual interface, yet port-hour and carrier cross-connect charges continue. Which teardown action is necessary?

Delete the virtual interface again until AWS automatically removes the underlying physical Direct Connect connection and every partner contract.
After safely removing remaining virtual interfaces, delete the provisioned Direct Connect connection; separately cancel the carrier and cross-connect arrangements as required.
Increase the BGP keepalive interval until the provider interprets the circuit as idle and stops monthly port and partner charges.
Change Route 53 DNS routing to VPN and assume it automatically terminates the still-provisioned AWS physical connection and third-party last-mile circuit.

Explanation

Correct Answer

AWS documents that deletion of the Direct Connect connection stops its port-hour charges after required VIF removal, but provider cross-connect and last-mile contracts may require independent cancellation.

Incorrect Answers

A virtual interface is a logical routing object. Deleting one does not necessarily delete the underlying AWS connection or cancel third-party connectivity contracts.

BGP control-plane timers affect session operation, not the customer commitment to a provisioned port or third-party circuit contract.

DNS selection has no authority to terminate an AWS Direct Connect physical connection or the separate service order maintained by the carrier.

Question 793

A security architecture routes east-west traffic from 80 spoke VPCs through a centralized inspection VPC attached to a Transit Gateway. Finance sees both attachment-hours and substantial traffic-processing usage even though each application request was sent once. What should the architect model?

Treat every inspected packet as free of Transit Gateway data-processing charges because inspection happens in another attached VPC.
Change the Transit Gateway route-table associations without changing any VPC attachment or inspected byte path and expect the hourly and usage charges to disappear.
Model attachment-hour charges and data processed on the actual routed hops, including inspection traffic, before choosing inspection placement and service scaling.
Remove only VPC Flow Logs while keeping every Transit Gateway attachment and inspected data path; assume log removal eliminates TGW processing charges.

Explanation

Correct Answer

Hub-and-spoke inspection can create extra routed traffic beyond application payloads. Cost assessment must map actual forwarding paths, attach resources and billing attribution, without removing required security controls.

Incorrect Answers

Transit Gateway is metered for network attachments and processed data regardless of whether a separate inspection appliance handles security policy.

Routing rules may optimize or segment paths, but an unchanged physical TGW attachment and byte volume still create the relevant hourly/traffic billing dimensions.

Flow logging is a separate telemetry service. Disabling it does not remove TGW attachments, suppress routed traffic bytes or waive TGW processing charges.

Question 794

A team deletes a cross-Region replication rule but leaves old destination copies; storage costs remain. Why?

Removing a replication rule automatically deletes every previously created destination object and cancels all accumulated replica storage.
Stopping future replication does not automatically delete existing destination objects already stored in the replica buckets.
Disable S3 bucket versioning in the source Region and assume destination replica objects previously copied to another Region disappear immediately.
Set the S3 replication rule priority to zero after deletion and assume priority changes retroactively reclaim destination storage.

Explanation

Correct Answer

Perform owner-approved destination lifecycle cleanup separately after safeguarding required recovery copies.

Incorrect Answers

Existing destination copies remain until a separate, policy-compliant deletion or lifecycle action removes them.

Source versioning changes do not automatically delete existing destination objects; dest lifecycle or explicit deletes may be needed.

Rule priority only resolves replication-rule precedence during future decisions; it does not erase prior replica objects.

Question 795

A low-latency single-Region application adds Global Accelerator and sees minimal user-visible improvement but new fees. Why?

Global Accelerator always lowers RTT by the same fixed percentage in every geography, independently of existing Internet paths.
Increase the Global Accelerator traffic dial in the same nearby Region and assume more endpoints necessarily reduce the Internet distance for current local users.
Enable accelerator static anycast addresses and assume the addresses alone provide an equal measured reduction in RTT for every user location.
There may be little path optimization or endpoint failover benefit for the measured client population to justify accelerator cost.

Explanation

Correct Answer

Choose based on observed performance and fixed-IP needs, not the existence of a global edge product.

Incorrect Answers

Real benefit varies by client population, routes and endpoints; benchmark before accepting additional accelerator charges.

A traffic dial changes endpoint routing allocation but does not guarantee lower latency for clients already on favorable paths.

Static anycast endpoint stability does not ensure measurable latency improvement over existing routes; test real users and cost.

Question 796

An application dashboard reports little EC2 NetworkOut, but AWS Data Exports shows significant inter-Region transfer charges attributed to the application account. The account also runs S3 replication and managed database replicas. Why can the metrics disagree?

EC2 NetworkOut covers only its instance-related interface metrics; S3 replication and managed services can generate separately metered transfer that requires billing/service reconciliation.
All managed S3 and database cross-Region replication bytes are included in the NetworkOut metric of whichever EC2 instance has the most CPU usage.
If an EC2 NetworkOut metric is low, AWS cannot correctly bill any S3 replication or managed database cross-Region traffic in the same account.
Change CloudWatch EC2 metrics to one-minute periods and assume this merges managed S3 replication charges into the historical instance network byte counter.

Explanation

Correct Answer

An instance network counter is not a consolidated bill across S3, database replication, NAT and other services. Reconcile Data Exports usage types with each service and the actual application cost tags/account ownership.

Incorrect Answers

EC2 instance metrics measure traffic on that instance, not data copied by S3 or a managed database service. CPU utilization is also not an estimator of independently billed replication bytes.

Managed-service transfer is metered by the service or account, so a low EC2 NetworkOut does not prove that the Data Exports line items are unrelated or incorrect.

More detailed EC2 metric intervals improve observation of that instance, but do not add independently metered managed-service traffic to its interface byte counter.

Question 797

Users near the target S3 Region see negligible upload improvement after enabling Transfer Acceleration, yet fees increase. Why?

Short/local paths may not gain meaningful benefit from accelerated edge routing while eligible accelerated bytes incur charges.
S3 Transfer Acceleration has no incremental charge on any path, so enabling it can never worsen cost economics.
Disable S3 Transfer Acceleration for all users without benchmarking distant clients and assume the lowest average latency is guaranteed for every client.
Use S3 multipart upload with more parallel requests but retain acceleration for nearby clients, assuming multipart concurrency removes the acceleration surcharge.

Explanation

Correct Answer

Use speed comparisons from representative clients to decide where acceleration makes economic sense.

Incorrect Answers

Acceleration pricing depends on eligible accelerated bytes and regions; local uploads may see little benefit relative to extra charges.

Disabling acceleration may save its fees but can worsen distant-client throughput; the near-Region cohort’s low benefit does not prove the global optimum.

Multipart improves retry/throughput mechanics but does not automatically waive accelerated request/data transfer pricing.

Question 798

A team selects CloudFront Price Class 100 and expects optimal low latency for viewers in all distant countries. Why is that not guaranteed?

Restricting CloudFront edge locations leaves every customer on the same nearest physical edge with unchanged latency and egress charges.
Restricting eligible edge locations can route distant viewers farther away than unrestricted delivery.
Use CloudFront Price Class 100 and assume CloudFront automatically adds every excluded edge location on demand for distant viewers.
Reduce the object TTL to zero and assume origin cache misses eliminate geographical latency differences for all viewers.

Explanation

Correct Answer

Compare geographical demand and performance rather than assuming the lowest price class also maximizes global performance.

Incorrect Answers

A reduced price class can alter eligible serving edges and increase latency for excluded geographies; compare experience and cost.

A restricted price class limits the eligible edge footprint and can route traffic to farther participating sites.

Cache TTL changes origin request patterns but do not restore excluded edge locations or guarantee a low-latency path.

Question 799

An operator moves every application instance and the only database into one AZ to remove inter-AZ charges; a single AZ outage takes down the entire service. What was sacrificed?

Same-AZ data placement automatically preserves multi-AZ failure protection after all standby replicas are deleted.
Multi-AZ fault tolerance was eliminated to optimize inter-AZ transfer fees.
Keep a single-AZ database but enable automatic storage snapshots and assume snapshots give synchronous Multi-AZ serving continuity.
Place primary and all standby instances in the same AZ but configure a cross-Region object-copy job as the application’s sole failover mechanism.

Explanation

Correct Answer

Retain independent failure domains and budget the cross-AZ traffic that real HA requires.

Incorrect Answers

Removing independent replicas changes the resilience architecture, even if it lowers normal cross-AZ transfer charges.

Backups permit recovery but do not supply a healthy serving replica in another AZ with automatic failover.

A copy job alone cannot make an entire same-AZ application/database stack survive immediate AZ loss with serving continuity.

Question 800

A VPC traffic mirror session copies production traffic to a central collector, and the network bill climbs despite unchanged app requests. Why?

Mirrored traffic reuses original app packets without creating any new transport or collector work, so enabling it cannot increase usage.
Limit mirror traffic to selected ports but assume the mirror target receives no new packets or collector work because the original application sends the same request volume.
Enable high-bandwidth VPC traffic mirroring without collector capacity planning and assume adding an extra traffic copy leaves billed data transfer unchanged.
Each mirrored packet produces extra traffic and collector processing beyond the original application flow.

Explanation

Correct Answer

Mirror filters and monitoring requirements must be weighed against incremental network costs.

Incorrect Answers

Traffic mirroring creates extra copied data and inspection processing, even when the application workload itself is unchanged.

Filtering may reduce copies but selected mirrored packets still traverse extra paths and consume collection resources.

Mirroring adds copied packet transport and inspection load independent of original application request counts.

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

Question 790

Explanation

Correct Answer

Incorrect Answers

Question 791

Explanation

Correct Answer

Incorrect Answers

Question 792

Explanation

Correct Answer

Incorrect Answers

Question 793

Explanation

Correct Answer

Incorrect Answers

Question 794

Explanation

Correct Answer

Incorrect Answers

Question 795

Explanation

Correct Answer

Incorrect Answers

Question 796

Explanation

Correct Answer

Incorrect Answers

Question 797

Explanation

Correct Answer

Incorrect Answers

Question 798

Explanation

Correct Answer

Incorrect Answers

Question 799

Explanation

Correct Answer

Incorrect Answers

Question 800

Explanation

Correct Answer

Incorrect Answers

AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests