Amazon AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 72

Question 710

A Spot instance is reclaimed with short notice and an uncheckpointed 12-hour job loses all progress. What design assumption failed?

The workload treated interruptible Spot capacity as guaranteed continuous compute without durable checkpoints.
Spot instances guarantee uninterrupted completion for the full 12-hour job whenever the purchase was placed during low-demand hours.
The Spot price fell during the job, which guarantees that any uncheckpointed application files will be recovered after capacity is reclaimed.
The batch worker had no second instance type configured, so the original reclaimed instance was obliged to finish its task uninterrupted.

Explanation

Correct Answer

Decouple job state from instance lifetime and use retry/checkpoint mechanisms or reserved capacity where necessary.

Incorrect Answers

Spot capacity can be reclaimed; the workload needs interruption-aware retries and durable progress even during off-peak periods.

Spot market pricing does not persist instance-local job state; interrupted jobs need durable progress and a restart path.

Capacity diversification improves replacement probability but cannot make an individual Spot instance interruption-proof; checkpoints remain necessary.

Question 711

A finance team buys a one-year EC2 commitment for a workload used 10 hours each quarter and observes low utilization of the commitment. Why?

An On-Demand EC2 instance can be reclaimed like Spot at any arbitrary moment solely to rebalance AWS spare capacity.
The committed hourly spend remains due regardless of whether the intermittent workload runs.
The one-year EC2 commitment is charged only for the ten running hours in each quarter and automatically becomes free on idle days.
Commitment utilization must always equal one hundred percent because purchases are prepaid, regardless of whether eligible EC2 usage exists.

Explanation

Correct Answer

Compare total break-even usage rather than assuming every discount commitment saves money.

Incorrect Answers

On-Demand does not have the same Spot interruption market behavior, making it suitable for brief non-interruptible use.

Hourly spending commitments persist throughout the term, even when qualifying usage is absent. Very intermittent work can leave the purchased benefit underutilized.

Prepayment does not create matched running usage. Utilization metrics identify idle commitment hours and whether the baseline justified the purchase.

Question 712

A team downsizes EC2 from CPU charts alone and experiences OOM kills despite low CPU utilization. What was missing?

Memory working-set and peak-memory requirements were not included in rightsizing analysis.
Enable EBS gp3 volume autoscaling and assume that increasing the attached disk size also increases the smaller instance class RAM enough to avoid OOM.
CPU utilization below twenty percent proves an instance cannot exhaust memory during any request burst.
A smaller EC2 instance keeps the original memory capacity as long as its attached EBS volume has enough free disk space.

Explanation

Correct Answer

CPU averages cannot prove a smaller instance has sufficient RAM, network or I/O capacity.

Incorrect Answers

EBS block volume growth is independent of the EC2 instance RAM allocation. Memory working sets must fit the selected compute instance type, regardless of available disk space.

Memory pressure is independent of CPU utilization; working-set/peak RSS can exceed smaller instance capacity despite low CPU.

EBS volume space is block storage, not instance RAM. Downsizing reduces available memory according to the chosen instance class.

Question 713

A binary-only x86 application is redeployed unchanged on Graviton and fails to start despite attractive pricing. Why?

Any x86-compiled program runs unchanged on Graviton because both families are EC2 instances under the same IAM permissions.
Use an ARM64-compatible container base image but retain an x86-only closed-source native library within the image, assuming the base-image architecture alone ensures the application starts.
The application must enable an x86-only EC2 Instance Savings Plan before the Graviton operating system can start an ARM64 executable.
The executable and native dependencies are not compiled for the ARM64 instruction set used by Graviton.

Explanation

Correct Answer

Use multi-architecture images or rebuild dependencies before trusting projected savings.

Incorrect Answers

Processor architecture affects executable and native library compatibility; IAM and tenancy similarity does not translate instruction sets.

An ARM64-compatible operating system layer does not translate the x86 machine code in a closed-source binary dependency. All native executables and libraries used by the process must be compatible.

Purchase commitments affect billing, not CPU instruction set compatibility. Use a compatible operating system image and compiled dependencies.

Question 714

An Auto Scaling group remains at desired capacity 200 after traffic falls because its minSize is set to 200. Why?

Auto Scaling decreases instance count below the group minimum simply because a CPU target-tracking policy requests less load.
The Auto Scaling group minimum capacity prohibits automatic scale-in below 200 even when metrics justify fewer instances.
The scale-in failure is caused by all Auto Scaling groups requiring a minimum of 200 instances regardless of configuration.
Maximum capacity equal to two hundred forces the group to keep two hundred instances running even when its minimum is lower.

Explanation

Correct Answer

Review minimum and desired capacity bounds and scale-in policies together.

Incorrect Answers

Group minimum capacity is an enforced lower bound; it must be configured consistently with the expected low-traffic floor.

Auto Scaling group minimum size is configurable and can be reduced when service objectives allow. This specific group is held at 200 by its own setting.

Maximum is an upper bound, not a required running count. The minimum and scaling policies determine the floor; here the minimum itself is 200.

Question 715

An EC2 Auto Scaling group has minimum size 1 and desired size 20. A scheduled action reduces desired size to 4, but the Activity history says it cannot reach desired capacity because every remaining instance is protected from scale-in. What must the operators address to allow safe scale-in?

Increase the group minimum size to 20 and rerun the scale-in action.
Enable EC2 API termination protection on all instances and assume it clears the Auto Scaling protected-instance setting.
Identify instances safe to terminate and remove EC2 Auto Scaling instance scale-in protection from enough instances, after draining their workloads.
Raise the Auto Scaling group maximum size and expect the protected instances to become eligible for automatic termination.

Explanation

Correct Answer

The Activity history explicitly identifies protected instances. After validating in-flight work and draining where needed, clear scale-in protection on eligible instances so Auto Scaling can terminate them toward the desired count.

Incorrect Answers

A higher minimum makes scale-in less possible and does not remove protection. The requested desired count is already above the configured minimum of one.

Auto Scaling instance scale-in protection is a separate control that decides which in-service instances may be terminated during scale-in. EC2 API termination protection does not disable it.

The maximum size is a ceiling for scale-out, not a control that releases instances from scale-in protection. Changing it cannot solve the reported protected-instance blocker.

Question 716

An event-processing application invokes Lambda many millions of times per day with long, near-continuous concurrent executions. Its metered request and duration bill exceeds an equivalently performing, well-utilized container fleet. Why can this happen?

Lambda has a mandatory 24-hour instance reservation fee even when it receives no invocations.
Lambda request-based billing stops charging execution duration once aggregate invocation volume becomes high enough.
Per-request and GB-second pricing can become uneconomical for high sustained duty cycles versus continuously utilized provisioned capacity.
Any Lambda function with more than one million invocations per month is automatically converted to a lower flat EC2 instance-hour rate.

Explanation

Correct Answer

Compare normalized cost per successful request at real load rather than treating serverless as always cheaper.

Incorrect Answers

Lambda generally charges for requests and execution resource use, subject to plan/pricing details, not perpetual EC2-style idle compute.

Lambda still charges eligible execution duration and requests at high sustained load; a continually busy fleet may make always-on alternatives economical.

Lambda and provisioned container/instance economics are separate billing models; there is no automatic wholesale conversion based on request count.

Question 717

An ECS Fargate service reserves 8 vCPU and 32 GiB per task but uses 1 vCPU and 3 GiB; costs remain high. Why?

Fargate bills on configured task resources rather than observed utilization inside an oversized task.
Fargate task charges depend only on observed CPU utilization after launch, so overallocating task vCPU has no effect on price.
Reduce the service desired task count while keeping the one remaining Fargate task overprovisioned at 8 vCPU and 32 GiB, assuming this automatically lowers cost per running task to 1 vCPU and 3 GiB.
Changing the ECS task to use fewer CPU/memory resources cannot lower its Fargate charge unless the number of cluster EC2 nodes also falls.

Explanation

Correct Answer

Benchmark right-sized task definitions and task counts; managed execution does not eliminate overprovisioning.

Incorrect Answers

Fargate billing follows configured task resources and duration, not a free right to reserve excess vCPU/RAM.

Fewer tasks may reduce total service charges but the remaining task is still billed for its configured 8-vCPU/32-GiB resource allocation. Rightsize its task definition separately.

ECS Fargate does not require managing EC2 host capacity for each task; right-sizing task CPU/memory can reduce its directly metered resource cost.

Question 718

A team stops an EC2 instance normally and expects the next boot to resume the unsaved browser session. What did it confuse?

Any EC2 instance can be hibernated even if its AMI, root EBS configuration and instance type do not support hibernation.
A standard stop/start is not hibernation; process memory was not saved for restoration.
A normal EC2 stop persists all unsaved RAM process state in the instance root EBS volume by default.
Restore a snapshot of the EBS boot volume on another EC2 instance and assume the unsaved browser process resumes with all in-memory tabs from the old normal-stop session.

Explanation

Correct Answer

Verify hibernation was supported, enabled and actually requested.

Incorrect Answers

EC2 hibernation requires eligible instance/AMI/configuration prerequisites; standard stop does not automatically save RAM.

Ordinary stop preserves persistent disks but not application memory. Hibernation is a separate configured feature that can persist RAM on supported instances.

Restoring persisted disk state does not recover process memory lost during a normal shutdown. Supported EC2 hibernation saves RAM explicitly before stopping, unlike ordinary EBS volume snapshots.

Question 719

A database team purchased a long RDS reserved compute commitment for a cluster that stays idle most weeks. Bills remain high despite idle compute. Why?

Aurora Serverless v2 automatically pauses to zero in every engine, cluster configuration and Region regardless of minimum ACU and eligibility.
The reserved-capacity commitment charges persist despite low actual utilization of the billed provisioned instances.
RDS Reserved Instance billing automatically stops on idle weekdays whenever the database serves no queries.
Pausing SQL statements on a provisioned RDS instance makes the purchased reserved-term charge refundable for unused hours.

Explanation

Correct Answer

Compare serverless capacity floors and supported engine constraints before committing to fixed-size billing.

Incorrect Answers

Auto-pause depends on supported engine/version and cluster settings, including configured minimum capacity, not a universal zero-floor guarantee.

A Reserved DB Instance is a term billing discount commitment tied to eligible capacity, not pay-per-query compute that disappears with idle demand.

The commitment is not refunded when the workload is quiet. Compare utilization assumptions and eligible savings against more elastic database choices.

Question 720

A team adopts a Compute Optimizer downsizing recommendation without testing a rare end-of-quarter workload and breaches its latency SLO. What was wrong?

A high-level EC2 CPU chart accurately represents every instance’s RAM and I/O demands for all time, even if peak workloads are missing.
Compute Optimizer guarantees that a historical utilization recommendation has already passed every future business-specific latency SLO.
Change the EC2 instance family on the recommendation without rechecking operating-system, CPU-architecture and native-library compatibility with the end-of-quarter batch.
Historical recommendation input did not establish performance under the unrepresented peak workload.

Explanation

Correct Answer

Recommendations should be validated against the target workload and memory/network/IO headroom.

Incorrect Answers

CPU alone is insufficient: representative memory, disk, network and seasonal load can make a smaller instance unsuitable.

Recommendations derive from observed metrics and assumptions, not unseen rare workloads; representative peak testing remains necessary.

Even if the historical utilization is representative, a family or architecture change needs application compatibility checks. The unobserved peak memory/I/O demand in this case also requires load testing.

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

Question 710

Explanation

Correct Answer

Incorrect Answers

Question 711

Explanation

Correct Answer

Incorrect Answers

Question 712

Explanation

Correct Answer

Incorrect Answers

Question 713

Explanation

Correct Answer

Incorrect Answers

Question 714

Explanation

Correct Answer

Incorrect Answers

Question 715

Explanation

Correct Answer

Incorrect Answers

Question 716

Explanation

Correct Answer

Incorrect Answers

Question 717

Explanation

Correct Answer

Incorrect Answers

Question 718

Explanation

Correct Answer

Incorrect Answers

Question 719

Explanation

Correct Answer

Incorrect Answers

Question 720

Explanation

Correct Answer

Incorrect Answers

AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests