Amazon AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 40

Question 390

An Aurora Global Database secondary has replication lag. A regional cutover immediately after a primary writer failure loses the last few writes, contrary to an assumed zero-RPO promise. What was overlooked?

Cross-Region Aurora Global Database replication is typically asynchronous; failover must account for lag and selected recovery procedure.
Aurora Global Database uses no replication between Regions.
Assume the secondary Aurora Region has every acknowledged primary transaction because its read-only cluster is queryable, without measuring replication lag.
Changing RDS instance tags makes replication synchronous by default.

Explanation

Correct Answer

Low lag does not imply that every unreplicated transaction will survive an unplanned regional failover.

Incorrect Answers

It does replicate; the issue is its timing and recovery semantics.

A readable secondary can still lag the primary during asynchronous cross-Region replication. Unplanned recovery may lose recent unreplicated changes.

Tags do not change data propagation guarantees.

Question 391

S3 Cross-Region Replication is enabled but recently uploaded objects do not appear in the backup bucket. The replication rule filters by prefix=archive/, while new keys start with records/. What caused the gap?

S3 Versioning always replicates every source object regardless of replication filters.
The object keys do not match the replication rule’s selected source filter.
Modify the destination bucket’s default replication prefix filter while leaving the source rule set to archive/, expecting records/ objects to match.
Bucket ownership enforcement disables all possible S3 replication for any object.

Explanation

Correct Answer

Replication only applies to qualifying objects under configured rules and permissions.

Incorrect Answers

Versioning does not override CRR rule scope.

S3 replication eligibility is evaluated by the configured source replication rule. Destination-side name changes cannot make records/ match archive/.

Bucket Owner Enforced is compatible with supported replication setups.

Question 392

A bucket has S3 Versioning enabled. A restore script issues GET on the key and receives 404 after a user deleted the object, although older data versions still exist. Why?

Versioning removes previous objects whenever a delete marker is inserted.
The current object version is a delete marker; retrieve a previous version or remove the marker through an authorized restore.
Enable S3 Versioning after the delete marker is already current and assume a plain GET will always skip the marker automatically.
A new Route 53 CNAME rebuilds the prior object version.

Explanation

Correct Answer

A simple GET may show a key as deleted while historical object versions remain stored.

Incorrect Answers

Versioning preserves earlier object versions subject to lifecycle/retention.

Versioning preserves historical revisions, but a current delete marker still masks the object key for an ordinary GET until a specific version is selected or the marker is removed.

DNS cannot restore object history.

Question 393

An AWS Backup plan creates protected recovery points in the primary Region but the secondary Region vault is always empty. The architect assumed selecting a second Region in documentation was sufficient. What configuration is missing?

A Route 53 health-check alarm automatically copies recovery points across Regions.
Backup plans always copy to every enabled AWS Region by default.
Create a destination AWS Backup vault and assume the mere existence of that vault starts asynchronous copies from every source backup plan.
A supported cross-Region backup copy action, destination vault and required KMS/IAM permissions.

Explanation

Correct Answer

A primary backup job does not independently create cross-region recovery points without copying them.

Incorrect Answers

DNS failover does not replicate backups.

Cross-region copies must be configured and supported for the resource type.

A destination vault is a target resource; a supported cross-region copy action and permissions must be configured before recovery points populate it.

Question 394

A company selects a pilot-light DR pattern but its runbook assumes a five-minute recovery while compute must be provisioned, configured and validated from zero. Which design assumption is flawed?

A pilot light generally needs scale-up and integration steps; the actual RTO must include those operations.
Pilot light always means full-sized application servers are running and serving traffic.
Recovery time excludes security and DNS validations by definition.
S3 Object Lock automatically provides an application’s five-minute RTO.

Explanation

Correct Answer

Low-cost minimal standby can increase restoration time; five-minute RTO requires measured deployment timing or warmer capacity.

Incorrect Answers

That is closer to warm standby or active-active, not minimal pilot light.

RTO measures the time until a usable recovered service, including dependencies.

Immutable storage does not launch an application stack.

Question 395

A warm-standby DR stack is running at 10% of normal capacity. A regional failover sends 100% of traffic immediately, causing overload. Which recovery assumption was invalid?

The standby environment still needs timely scale-out and capacity verification before serving full production demand.
A 10% standby automatically scales to 100% before any request regardless of quotas.
Route 53 DNS records guarantee adequate EC2 instance quotas.
Configure Route 53 rapid DNS failover but leave the warm standby Auto Scaling group capped at 10% of production traffic and no tested scale-out policy.

Explanation

Correct Answer

Warm standby reduces deployment effort but is not necessarily pre-provisioned for all traffic.

Incorrect Answers

Scaling depends on policy, available quota/capacity and time.

Routing does not supply compute capacity.

Quick DNS cutover can overload the reduced-size standby unless scaling capacity, regional quotas and readiness are included in the recovery sequence.

Question 396

An EC2 Auto Scaling group launches healthy instances but its ALB uses a health path that always returns 404 for newly deployed versions. Targets cycle in and out despite working root pages. Why?

Auto Scaling always ignores load balancer health checks.
Set ALB target-group matcher to accept HTTP 404 as healthy so all new instances pass even though the configured health endpoint does not exist.
Changing the RDS storage class fixes an ALB 404 health path.
Health check settings do not match the application’s supported health route/expected response and trigger unnecessary replacement.

Explanation

Correct Answer

Health policy must test a valid stable endpoint so the fleet is not marked unhealthy incorrectly.

Incorrect Answers

It can replace unhealthy instances based on configured health types.

Accepting 404 masks a bad probe path rather than checking application readiness. Correct the requested path or response codes to match an actual health endpoint.

Database storage does not correct the probe URI.

Question 397

Three private subnets each use a NAT gateway in us-east-1a. When us-east-1a has an outage, private EC2 instances in us-east-1b and us-east-1c cannot reach updates. Why?

All subnet routes depend on NAT in a single AZ, creating an avoidable shared egress failure domain.
Internet gateways must be installed once per AZ rather than per VPC.
Auto Scaling replaces NAT gateways in different AZs automatically without configuration.
Create NAT gateways in the application AZs but keep all AZ-b/c route tables pointing to the original AZ-a gateway, expecting location alone to correct routing.

Explanation

Correct Answer

Use zone-local NAT/route strategy or validated alternative as the required availability dictates.

Incorrect Answers

An internet gateway is VPC-scoped; NAT path placement is the issue.

Managed NAT gateways do not appear in every AZ merely because subnets exist.

Merely provisioning alternate NAT gateways does not remove the shared failure path. Private subnet route tables must actually target healthy AZ-local or supported regional NAT.

Question 398

An interface VPC endpoint exists only in subnet 10.2.1.0/24 in AZ-a. During that AZ outage, private clients in AZ-b cannot reach the service even though service Region remains healthy. What design gap exists?

The VPC lacks endpoint network interfaces in the other required AZs, leaving a single AZ endpoint path.
The service must always use the public internet because private endpoints cannot span AZs.
IAM permissions boundaries automatically create missing endpoint ENIs.
Change the interface endpoint security group to allow all protocols while leaving its only endpoint ENI in the AZ that is unavailable.

Explanation

Correct Answer

Multiple supported AZ endpoint ENIs improve availability; verify DNS resolution and routing/security-group paths.

Incorrect Answers

Interface endpoints can be provisioned in supported subnets across AZs.

IAM policies do not provision private network interfaces.

A permissive policy cannot create a data-plane endpoint interface in a surviving AZ. Add supported subnets/ENIs and maintain appropriate security restrictions.

Question 399

A web order API uses an asynchronous SQS queue but responds “completed” immediately. During inventory outages, paid orders sit queued for hours and customers complain about false fulfillment. What is wrong?

The response conflates durable queue acceptance with downstream business commit.
SQS cannot durably accept any events while consumers are offline.
Adding DNS caching turns queued work into completed inventory reservations.
Using an SQS FIFO queue would make all backend side effects happen synchronously.

Explanation

Correct Answer

Use pending/accepted status and report completion only on authoritative worker/transaction evidence.

Incorrect Answers

Queue buffering exists specifically to decouple producer and consumer availability.

DNS caching changes hostname lookups; it cannot execute an inventory transaction that remains in an asynchronous queue.

FIFO ordering does not change asynchronous completion boundaries.

Question 400

An SNS topic publishes a critical alert to two endpoints. The team subscribed only direct HTTP consumers with no durable queue, and a consumer was unavailable long enough to exhaust delivery retries. Why was its copy lost?

SNS topics inherently retain every message for arbitrary replay by all late subscribers.
A CloudFront cache TTL determines whether SNS messages survive an HTTP outage.
Set the SNS HTTP subscription delivery retry policy to only a few seconds and assume it becomes an unlimited durable independent backlog after that period.
The delivery path lacked a durable per-consumer buffer and exhausted its retry/delivery policy.

Explanation

Correct Answer

Separate SQS subscriptions can retain each consumer’s events while downstream processing is unavailable.

Incorrect Answers

SNS is fanout delivery, not an unlimited replayable stream.

CDN caching does not retain SNS delivery copies.

Direct HTTP delivery has bounded retries; configure a durable SQS subscription or supported delivery-failure route when offline retention is required.

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

Question 390

Explanation

Correct Answer

Incorrect Answers

Question 391

Explanation

Correct Answer

Incorrect Answers

Question 392

Explanation

Correct Answer

Incorrect Answers

Question 393

Explanation

Correct Answer

Incorrect Answers

Question 394

Explanation

Correct Answer

Incorrect Answers

Question 395

Explanation

Correct Answer

Incorrect Answers

Question 396

Explanation

Correct Answer

Incorrect Answers

Question 397

Explanation

Correct Answer

Incorrect Answers

Question 398

Explanation

Correct Answer

Incorrect Answers

Question 399

Explanation

Correct Answer

Incorrect Answers

Question 400

Explanation

Correct Answer

Incorrect Answers

AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests