cert
cert-1
cert-2

Pass Amazon AWS Certified Solutions Architect - Associate SAA-C03 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
AWS Certified Solutions Architect - Associate SAA-C03 Exam - Verified By Experts
AWS Certified Solutions Architect - Associate SAA-C03 Premium Bundle
$29.98

AWS Certified Solutions Architect - Associate SAA-C03 Premium Bundle

$79.99
$109.97
  • Premium File 1063 Questions & Answers. Last update: Nov 10, 2026
  • Training Course 368 Video Lectures
  • Study Guide 632 Pages
 
$109.97
$79.99
accept 217 downloads in last 7 days
block-screenshots
Purchase Amazon AWS Certified Solutions Architect - Associate SAA-C03 Exam Training Products Individually
Amazon AWS Certified Solutions Architect - Associate SAA-C03 Premium File
Premium File 1063 Q&A
$76.99$69.99
Amazon AWS Certified Solutions Architect - Associate SAA-C03 Video Training Course
Training Course 368 Lectures
$21.99 $19.99
Amazon AWS Certified Solutions Architect - Associate SAA-C03 PDF Study Guide
Study Guide 632 Pages
$21.99 $19.99
Amazon AWS Certified Solutions Architect - Associate SAA-C03 Practice Test 72

Question 710

A Lambda image-processing function uses maximum memory despite low utilization. What optimization controls both cost and latency?

  1. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration.
  2. Schedule provisioned concurrency only during justified low-latency demand periods.
  3. Use an appropriately sized Lambda function for eligible short sporadic executions.
  4. Benchmark alternative Lambda memory configurations using duration and cost together.

Explanation

Correct Answer

D:

Lambda CPU scales with memory, so lower memory may run longer; minimum GB-seconds is not always lowest cost. Compare total request charges and p95 duration across memory sizes.

Incorrect Answers

A:

Compute selection must satisfy execution limits before cost comparison. Benchmark alternative Lambda memory configurations using duration and cost together. Compare total request charges and p95 duration across memory sizes.

B:

Provisioned concurrency has a time-based cost even when requests are absent. Benchmark alternative Lambda memory configurations using duration and cost together. Compare total request charges and p95 duration across memory sizes.

C:

Lambda pricing follows invocation and duration rather than always-on instance-hours. Benchmark alternative Lambda memory configurations using duration and cost together. Compare total request charges and p95 duration across memory sizes.

Question 711

A small web API has steady 24/7 traffic and substantial Lambda invocation cost. What compute alternative should be priced?

  1. Use an appropriately sized Lambda function for eligible short sporadic executions.
  2. Schedule one-off Fargate tasks and remove idle service replicas.
  3. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration.
  4. Compare ECS Fargate or EC2 reserved baseline against Lambda at actual sustained request levels.

Explanation

Correct Answer

D:

Persistent load can make provisioned container capacity economical, depending on operational overhead. Model annual cost at measured concurrency and request duration.

Incorrect Answers

A:

Lambda pricing follows invocation and duration rather than always-on instance-hours. Compare ECS Fargate or EC2 reserved baseline against Lambda at actual sustained request levels. Model annual cost at measured concurrency and request duration.

B:

Paying for active tasks only during job execution can eliminate unnecessary always-on task-hours. Compare ECS Fargate or EC2 reserved baseline against Lambda at actual sustained request levels. Model annual cost at measured concurrency and request duration.

C:

Compute selection must satisfy execution limits before cost comparison. Compare ECS Fargate or EC2 reserved baseline against Lambda at actual sustained request levels. Model annual cost at measured concurrency and request duration.

Question 712

A batch job runs a Fargate task every hour for five minutes but a full-time ECS service stays idle between runs. What should change?

  1. Schedule provisioned concurrency only during justified low-latency demand periods.
  2. Use an appropriately sized Lambda function for eligible short sporadic executions.
  3. Schedule one-off Fargate tasks and remove idle service replicas.
  4. Compare ECS Fargate or EC2 reserved baseline against Lambda at actual sustained request levels.

Explanation

Correct Answer

C:

Paying for active tasks only during job execution can eliminate unnecessary always-on task-hours. Inspect task lifetimes and billed vCPU/memory usage.

Incorrect Answers

A:

Provisioned concurrency has a time-based cost even when requests are absent. Schedule one-off Fargate tasks and remove idle service replicas. Inspect task lifetimes and billed vCPU/memory usage.

B:

Lambda pricing follows invocation and duration rather than always-on instance-hours. Schedule one-off Fargate tasks and remove idle service replicas. Inspect task lifetimes and billed vCPU/memory usage.

D:

Persistent load can make provisioned container capacity economical, depending on operational overhead. Schedule one-off Fargate tasks and remove idle service replicas. Inspect task lifetimes and billed vCPU/memory usage.

Question 713

A high-volume Lambda pipeline writes excessive CloudWatch debug logs, dominating cost. What change reduces secondary service charges?

  1. Reduce unnecessary log verbosity and apply justified CloudWatch Logs retention settings.
  2. Schedule provisioned concurrency only during justified low-latency demand periods.
  3. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration.
  4. Use an appropriately sized Lambda function for eligible short sporadic executions.

Explanation

Correct Answer

A:

Logging volume and retained bytes are billed independently of Lambda compute charges. Compare log-ingestion GB and retention after tuning.

Incorrect Answers

B:

Provisioned concurrency has a time-based cost even when requests are absent. Reduce unnecessary log verbosity and apply justified CloudWatch Logs retention settings. Compare log-ingestion GB and retention after tuning.

C:

Compute selection must satisfy execution limits before cost comparison. Reduce unnecessary log verbosity and apply justified CloudWatch Logs retention settings. Compare log-ingestion GB and retention after tuning.

D:

Lambda pricing follows invocation and duration rather than always-on instance-hours. Reduce unnecessary log verbosity and apply justified CloudWatch Logs retention settings. Compare log-ingestion GB and retention after tuning.

Question 714

A production Lambda has provisioned concurrency for all hours but receives traffic only during weekday peaks. Which pricing optimization fits?

  1. Schedule one-off Fargate tasks and remove idle service replicas.
  2. Benchmark alternative Lambda memory configurations using duration and cost together.
  3. Schedule provisioned concurrency only during justified low-latency demand periods.
  4. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration.

Explanation

Correct Answer

C:

Provisioned concurrency has a time-based cost even when requests are absent. Compare available concurrency, cold-start rate and monthly provisioned hours.

Incorrect Answers

A:

Paying for active tasks only during job execution can eliminate unnecessary always-on task-hours. Schedule provisioned concurrency only during justified low-latency demand periods. Compare available concurrency, cold-start rate and monthly provisioned hours.

B:

Lambda CPU scales with memory, so lower memory may run longer; minimum GB-seconds is not always lowest cost. Schedule provisioned concurrency only during justified low-latency demand periods. Compare available concurrency, cold-start rate and monthly provisioned hours.

D:

Compute selection must satisfy execution limits before cost comparison. Schedule provisioned concurrency only during justified low-latency demand periods. Compare available concurrency, cold-start rate and monthly provisioned hours.

Question 715

A retrying Lambda function repeatedly calls a failing third-party API, increasing invocation charges. What design improves cost and reliability?

  1. Schedule one-off Fargate tasks and remove idle service replicas.
  2. Use bounded exponential backoff, circuit breaking and DLQ or failure destination when appropriate.
  3. Reduce unnecessary log verbosity and apply justified CloudWatch Logs retention settings.
  4. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration.

Explanation

Correct Answer

B:

Runaway retries consume compute and can amplify an outage; bounded recovery reduces wasted calls. Force endpoint failure and inspect retry counts and total billed invocations.

Incorrect Answers

A:

Paying for active tasks only during job execution can eliminate unnecessary always-on task-hours. Use bounded exponential backoff, circuit breaking and DLQ or failure destination when appropriate. Force endpoint failure and inspect retry counts and total billed invocations.

C:

Logging volume and retained bytes are billed independently of Lambda compute charges. Use bounded exponential backoff, circuit breaking and DLQ or failure destination when appropriate. Force endpoint failure and inspect retry counts and total billed invocations.

D:

Compute selection must satisfy execution limits before cost comparison. Use bounded exponential backoff, circuit breaking and DLQ or failure destination when appropriate. Force endpoint failure and inspect retry counts and total billed invocations.

Question 716

A team proposes Lambda for an hours-long CPU-heavy task and assumes it is cheaper because it is serverless. What constraint invalidates the design?

  1. Benchmark alternative Lambda memory configurations using duration and cost together.
  2. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration.
  3. Schedule provisioned concurrency only during justified low-latency demand periods.
  4. Compare ECS Fargate or EC2 reserved baseline against Lambda at actual sustained request levels.

Explanation

Correct Answer

B:

Compute selection must satisfy execution limits before cost comparison. Run a realistic job length and confirm it completes without forced split or timeout.

Incorrect Answers

A:

Lambda CPU scales with memory, so lower memory may run longer; minimum GB-seconds is not always lowest cost. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration. Run a realistic job length and confirm it completes without forced split or timeout.

C:

Provisioned concurrency has a time-based cost even when requests are absent. Use an eligible long-running container or batch platform because Lambda has a maximum execution duration. Run a realistic job length and confirm it completes without forced split or timeout.

D:

Persistent load can make provisioned container capacity economical, depending on operational overhead. Compute selection must satisfy execution limits before cost comparison.

Question 717

A batch queue runs thousands of retryable jobs and must reduce EC2 cost without requiring guaranteed continuous capacity. Which configuration fits?

  1. Submit jobs to AWS Batch using eligible Spot or scheduled compute that scales to zero when idle.
  2. List multiple compatible instance types with a capacity-aware allocation strategy.
  3. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools.
  4. Retain an On-Demand minimum capacity or another stable baseline for noninterruptible service.

Explanation

Correct Answer

C:

Interruptible spare capacity can reduce cost when jobs resume safely after reclaim. Simulate Spot interruption and inspect completed-job rate.

Incorrect Answers

A:

Queued asynchronous jobs can share transient compute rather than permanent single-purpose servers. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools. Simulate Spot interruption and inspect completed-job rate.

B:

Diverse Spot pools improve odds of obtaining spare capacity without overpaying for one type. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools. Simulate Spot interruption and inspect completed-job rate.

D:

Spot-only fleets can vanish; a critical serving tier needs resilient capacity choices. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools. Simulate Spot interruption and inspect completed-job rate.

Question 718

An application uses 10 On-Demand servers at low traffic but could reduce to two outside peak hours. What should control fleet size?

  1. Retain an On-Demand minimum capacity or another stable baseline for noninterruptible service.
  2. Use instance utilization and request/SLO metrics to reset the justified minimum capacity.
  3. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools.
  4. Set a safe minimum and dynamic Auto Scaling target tracking based on a meaningful metric.

Explanation

Correct Answer

D:

Elastic desired capacity avoids paying for peak capacity during long quiet periods. Compare instance-hours and p95 service latency across a day.

Incorrect Answers

A:

Spot-only fleets can vanish; a critical serving tier needs resilient capacity choices. Set a safe minimum and dynamic Auto Scaling target tracking based on a meaningful metric. Compare instance-hours and p95 service latency across a day.

B:

A continually growing baseline can cause hidden idle fleet spending. Set a safe minimum and dynamic Auto Scaling target tracking based on a meaningful metric. Compare instance-hours and p95 service latency across a day.

C:

Interruptible spare capacity can reduce cost when jobs resume safely after reclaim. Set a safe minimum and dynamic Auto Scaling target tracking based on a meaningful metric. Compare instance-hours and p95 service latency across a day.

Question 719

Spot interruptions repeatedly remove the last available app instance, causing outages. What should be changed?

  1. Use instance utilization and request/SLO metrics to reset the justified minimum capacity.
  2. List multiple compatible instance types with a capacity-aware allocation strategy.
  3. Retain an On-Demand minimum capacity or another stable baseline for noninterruptible service.
  4. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools.

Explanation

Correct Answer

C:

Spot-only fleets can vanish; a critical serving tier needs resilient capacity choices. Reclaim every Spot instance and verify the service stays available.

Incorrect Answers

A:

A continually growing baseline can cause hidden idle fleet spending. Retain an On-Demand minimum capacity or another stable baseline for noninterruptible service. Reclaim every Spot instance and verify the service stays available.

B:

Diverse Spot pools improve odds of obtaining spare capacity without overpaying for one type. Retain an On-Demand minimum capacity or another stable baseline for noninterruptible service. Reclaim every Spot instance and verify the service stays available.

D:

Interruptible spare capacity can reduce cost when jobs resume safely after reclaim. Retain an On-Demand minimum capacity or another stable baseline for noninterruptible service. Reclaim every Spot instance and verify the service stays available.

Question 720

An Auto Scaling group always selects a single instance family despite unpredictable Spot supply. How should it improve purchase flexibility?

  1. List multiple compatible instance types with a capacity-aware allocation strategy.
  2. Use instance utilization and request/SLO metrics to reset the justified minimum capacity.
  3. Use a Spot-enabled Auto Scaling or AWS Batch compute environment with diverse capacity pools.
  4. Set a safe minimum and dynamic Auto Scaling target tracking based on a meaningful metric.

Explanation

Correct Answer

A:

Diverse Spot pools improve odds of obtaining spare capacity without overpaying for one type. Disable one pool and inspect replacement capacity.

Incorrect Answers

B:

A continually growing baseline can cause hidden idle fleet spending. List multiple compatible instance types with a capacity-aware allocation strategy. Disable one pool and inspect replacement capacity.

C:

Interruptible spare capacity can reduce cost when jobs resume safely after reclaim. List multiple compatible instance types with a capacity-aware allocation strategy. Disable one pool and inspect replacement capacity.

D:

Elastic desired capacity avoids paying for peak capacity during long quiet periods. Diverse Spot pools improve odds of obtaining spare capacity without overpaying for one type.

AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests