Amazon AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 45

Question 440

During an AWS Fault Injection Service experiment, a CloudWatch alarm for application checkout errors enters ALARM and the experiment stops before additional instance disruptions. What did that configured stop condition achieve?

It proved that checkout never experienced a single failed request during the controlled experiment.
It automatically restored every impacted EC2 and database resource without needing service recovery.
It guaranteed a passing business RTO solely because a monitoring alarm exists.
It acted as a safety guardrail to halt further fault injection when measured impact exceeded the approved threshold; the team must still inspect outcomes and recovery evidence.

Explanation

Correct Answer

AWS FIS stop conditions can be bound to alarms so an experiment ends when an approved impact threshold triggers, limiting further injected disruption.

Incorrect Answers

A stop condition can activate after an error threshold; it does not establish that no initial errors occurred.

Stopping future FIS actions is not identical to full rollback of every fault or recovery of all affected components.

An alarm is a blast-radius bound, while demonstrated RTO requires timing and application availability evidence.

Question 441

A DR launch test in Region B now passes its vCPU SERVICE QUOTA check, yet it repeatedly fails to launch the required instance family in AZ-2 because that specific AZ has insufficient On-Demand capacity. Which additional preparation is appropriate when guaranteed DR capacity is justified?

Request another increase to a quota that is already high enough and assume quotas guarantee physical AZ supply.
Change Route 53 DNS records to AZ-2 before verifying instance capacity and assume client traffic allocates free servers.
Evaluate available matching EC2 On-Demand Capacity Reservations or alternative tested instance/AZ capacity plans for the DR workload and retest recovery at scale.
Delete AWS Backup recovery points in Region A to free physical EC2 hardware in Region B.

Explanation

Correct Answer

AWS On-Demand Capacity Reservations can reserve matching instance type/platform/AZ/tenancy capacity subject to availability and cost. An approved alternative family or AZ can be a fallback.

Incorrect Answers

A service quota is an account ceiling, not a reservation of physical hosts in a particular Availability Zone.

DNS request routing cannot create EC2 host inventory or guarantee an instance type in an exhausted AZ.

Backup copies consume storage and are not a substitute for target-AZ EC2 compute supply.

Question 442

An ALB target group uses a deep health endpoint that fails whenever one shared database briefly disconnects. During a database incident ALL web targets become unhealthy simultaneously, yet new requests still reach them. What load-balancer behavior and readiness tradeoff explain the result?

ALB guarantees it drops every new HTTP request whenever all registered targets are unhealthy.
Turning off all target health checks provides a more reliable signal of application readiness during a shared dependency outage.
ALB can fail open and route to all unhealthy registered targets when none are healthy; a shared dependency in every readiness check can therefore amplify a common outage, so choose meaningful but stable health criteria.
An IAM identity-center permission set changes target-group fail-open routing based on database authorization.

Explanation

Correct Answer

Deep common dependency checks can mark every target unhealthy together; AWS ALB can then fail open. Separating local readiness from shared upstream circuit-breaker/recovery signals may limit cascading failure.

Incorrect Answers

AWS documents ALB fail-open to registered unhealthy targets when a target group has no healthy target. There is no guarantee all client requests are dropped.

Health checks need to reflect service health but disabling them removes traffic isolation and cannot address the common database failure.

Identity permission assignment does not govern ALB response when all target health checks fail.

Question 443

An Aurora cluster uses multi-AZ distributed storage with only one compute writer. Regional failure is now in scope. Which additional architecture addresses that larger failure boundary?

Assume six storage copies across three AZs also means global cross-region copies exist.
Use only more replicas in the same AZ as the current writer.
Add two additional Aurora reader instances in new AZs within the same Region and call this full regional disaster recovery.
Design a supported cross-Region Aurora replication/global-database and recovery procedure, including RPO/RTO tests.

Explanation

Correct Answer

Aurora’s multi-AZ cluster volume addresses AZ failures but does not itself replicate a database to another Region.

Incorrect Answers

AZ copies remain within the same Region.

Same-zone capacity cannot cover regional unavailability.

AZ replica compute and distributed storage improve regional AZ availability; a full regional outage requires data and recovery capacity in another Region.

Question 444

An RDS DB instance is scheduled for deletion. Finance needs to retain a restorable historical dataset for one year even after the live instance and its ordinary automated backup retention lifecycle end. What should be retained under the approved policy?

Create and retain an appropriate manual final DB snapshot with the required encryption/key access and tested restoration, since manual snapshots persist until explicitly deleted.
Assume the Multi-AZ standby continues to exist and preserve old business rows indefinitely after the DB instance is deleted.
Delete all manual snapshots because the deleted primary DB identifier itself stores historical backups in DNS.
Use Route 53 geolocation records as the permanent point-in-time historical backup.

Explanation

Correct Answer

Manual DB snapshots can outlive the source instance until intentional deletion; backup recovery still requires suitable encryption and a tested restore path.

Incorrect Answers

A deleted deployment has no continuing synchronous standby, and HA replication is not a historical archive.

Names and identifiers alone contain no restorable database blocks; deleting the retained snapshots can eliminate the required evidence.

DNS policies neither retain database history nor preserve encrypted snapshot bytes.

Question 445

After distributing ALB web targets across three AZs, how can the SRE team verify the AZ failure design rather than merely reading subnet assignments?

Count three subnet IDs but send no traffic during an outage.
Count ALB target registration and AZ IDs only without suppressing a test target AZ or measuring actual HTTP service after removal.
Disable health checks before the test to preserve all bad targets.
Run an approved one-AZ target outage test, check surviving target health and measure real client error/latency against SLO.

Explanation

Correct Answer

Actual failure testing demonstrates capacity in remaining zones and the ALB’s routing behavior.

Incorrect Answers

Placement inventory alone does not prove survivability.

Placement inventory is necessary but insufficient to prove that remaining zones have capacity and clients meet the SLO during a real failure.

That interferes with fault isolation.

Question 446

An RDS Multi-AZ DB instance has completed an approved failover test. Which evidence shows both automated takeover and that application connections followed the new writer?

Query the standard standby before failover and expect it to accept regular reporting SELECTs.
Check that the RDS endpoint hostname string did not change, but never reconnect clients or issue a post-failover write transaction.
Validate the old primary’s private IP remains the write address forever.
Inspect RDS failover events and reconnect to the stable DB endpoint, validating writes after promotion and measuring disruption.

Explanation

Correct Answer

An event log plus actual transaction checks proves the new primary is usable.

Incorrect Answers

The conventional standby is not an app reader.

A stable DNS name can resolve to a promoted writer but old client sessions and cached IPs may fail. Validate actual new connections and writes after promotion.

The writer may move after failover.

Question 447

An RDS Multi-AZ DB cluster claims read scale without compromising write routing. What end-to-end test confirms the deployed endpoint separation?

Send controlled writes to the writer endpoint and SELECTs to the reader endpoint, observing reader load and read-only enforcement.
Submit database writes to the read-only reader endpoint as proof of full read/write capability.
Point both the SQL write and reporting read clients at the writer endpoint and assume a Multi-AZ DB cluster balances the reporting sessions to readers automatically.
Check that one Multi-AZ standby cannot be queried and conclude all cluster readers are unavailable.

Explanation

Correct Answer

The two endpoints should serve intended roles on the actual supported DB cluster.

Incorrect Answers

Reader endpoint sessions are read-only.

Eligible read connections must target the reader endpoint; queries sent to the writer endpoint still consume writer capacity.

DB instance and DB cluster Multi-AZ models differ.

Question 448

An Aurora cluster has reader promotion tiers. Which drill confirms the selection behaves acceptably when a preferred reader is unavailable?

Verify only that every replica has the preferred name in its identifier.
Remove all readers before a failover and assume service continuity is guaranteed.
Compare only Aurora reader instance names after failover while ignoring writer-role evidence, events and whether the preferred candidate was unhealthy.
Simulate an approved writer failure with one preferred candidate removed, inspect the promoted writer, Aurora events and application reconnects.

Explanation

Correct Answer

Testing fallback eligibility verifies actual promotion behavior, not just tier configuration.

Incorrect Answers

Names do not determine healthy promotion outcomes.

Without an eligible reader promotion may require greater recovery work.

Instance names are not authoritative promotion evidence. Inspect promotion events, selected writer and client recovery behavior with eligibility constraints.

Question 449

An EC2 database is provisioned with gp3 EBS; random I/O saturation is observed, while storage space is ample. The team needs more IOPS without buying a proportionally larger volume. What should it do?

Increase only logical volume size until it automatically earns more gp3 baseline IOPS.
Provision higher gp3 IOPS and, if needed, independent throughput within the supported limits and EC2 bandwidth.
Switch to sc1 cold HDD for high-IOPS random transactions.
Enable ENA enhanced networking on the EC2 instance but leave its saturated gp3 provisioned IOPS unchanged, assuming the network adapter setting increases the attached volume IOPS allowance.

Explanation

Correct Answer

Unlike gp2, gp3 IOPS and throughput have provisioned parameters largely independent of volume size.

Incorrect Answers

gp3 performance is not scaled by storage size in the gp2 manner.

Cold HDD volumes are unsuitable for sustained random low-latency operations.

ENA is relevant to instance network throughput and latency, not the gp3 volume’s separately configured IOPS. When storage I/O is the proven bottleneck, provision supported volume performance and check instance EBS limits.

Question 450

An EC2 OLTP engine requires sustained high random IOPS and tight sub-millisecond storage latency beyond the desired gp3 tier. Which EBS volume family is appropriate if instance support is verified?

st1 HDD, which prioritizes sequential throughput rather than random IOPS.
Use a gp3 SSD with an IOPS setting that is already at its supported limit and assume a lower-priced storage class must meet any higher random-IOPS/latency requirement.
Provisioned IOPS io2 Block Express with suitable IOPS/throughput and instance compatibility.
sc1 Cold HDD with the same IOPS objective as io2.

Explanation

Correct Answer

io2 Block Express targets demanding transactional performance and low storage-latency distribution.

Incorrect Answers

HDD seek behavior does not meet low-latency random I/O requirements.

gp3 offers a configurable performance range, but workloads exceeding its supported limits may need provisioned IOPS io2 Block Express with a compatible instance and tested latency characteristics.

Cold HDD optimizes rarely accessed sequential data, not high transactional IOPS.

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

Question 440

Explanation

Correct Answer

Incorrect Answers

Question 441

Explanation

Correct Answer

Incorrect Answers

Question 442

Explanation

Correct Answer

Incorrect Answers

Question 443

Explanation

Correct Answer

Incorrect Answers

Question 444

Explanation

Correct Answer

Incorrect Answers

Question 445

Explanation

Correct Answer

Incorrect Answers

Question 446

Explanation

Correct Answer

Incorrect Answers

Question 447

Explanation

Correct Answer

Incorrect Answers

Question 448

Explanation

Correct Answer

Incorrect Answers

Question 449

Explanation

Correct Answer

Incorrect Answers

Question 450

Explanation

Correct Answer

Incorrect Answers

AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests