How to Reduce Lambda Compute Costs While Maintaining Latency?

SAA-C03 By islever Editorial Last updated

Design cost-optimized compute solutions. Design high-performing and elastic compute solutions.

Short answer

D Provisioned concurrency on the Lambda functions plus a Compute Optimizer memory increase shortens CPU-intensive duration and sustains latency.

Question

A company uses high concurrency AWS Lambda functions to process a constantly increasing number of messages in a message queue during marketing events. The Lambda functions use CPU intensive code to process the messages. The company wants to reduce the compute costs and to maintain service latency for its customers. Which solution will meet these requirements?
  1. Configure reserved concurrency for the Lambda functions. Decrease the memory allocated to the Lambda functions.
  2. Configure reserved concurrency for the Lambda functions. Increase the memory according to AWS Compute Optimizer recommendations.
  3. Configure provisioned concurrency for the Lambda functions. Decrease the memory allocated to the Lambda functions.
  4. Configure provisioned concurrency for the Lambda functions. Increase the memory according to AWS Compute Optimizer recommendations. Correct Answer

Key Takeaways

It tests whether you can separate Lambda's two concurrency controls — provisioned concurrency for cold-start latency and reserved concurrency for capping/guaranteeing a concurrency budget — while knowing that CPU scales with allocated memory.

This SAA-C03 question asks how to cut AWS Lambda compute costs for CPU-intensive message-queue processing while keeping customer-facing latency stable during marketing spikes. The page establishes that the answer is provisioned concurrency paired with a memory increase (D), not reserved concurrency.

Common mistake: Picking B (reserved concurrency) because it is free and appears to save money. Reserved concurrency only reschedules the account's concurrency pool; it does not pre-initialize execution environments, so cold starts remain and a marketing-event burst can be throttled instead of served.

Expert Analysis

Why the Answer Is Correct

Option D combines the two levers that directly address the stated requirements. Provisioned concurrency keeps execution environments initialized and ready, which is the documented remedy for cold-start latency on a workload whose message volume is, in Salilgen's words, "constantly increasing". Increasing memory matters because AWS Lambda allocates CPU proportionally to memory, so CPU-intensive code finishes sooner; since you are billed per GB-second, a shorter duration can offset the higher per-millisecond rate. LeonSauveterre made exactly this point, noting that "CPU resources scale proportionally with memory allocation in AWS Lambda", and Sivaeas reached the same conclusion about memory tuning for CPU-bound work. Together, provisioned concurrency preserves latency while extra memory shrinks duration-driven cost.

Why the Other Options Are Wrong

Option A and Option C both decrease memory, which reduces the available vCPU for CPU-intensive code and makes each invocation run longer — the opposite of cost reduction and a latency risk. Option B keeps reserved concurrency, which is a free feature but only partitions the account's concurrency pool; it guarantees or caps how many instances a function may run and does nothing to remove cold starts. Worse, a hard reserved-concurrency ceiling can throttle invocations precisely during a marketing-event burst, hurting the latency the company wants to protect. JonJon03 argued correctly that "There is no charge for configuring reserved concurrency for a function", but cost is only half the requirement — the latency half is unmet.

Community Comment Notes

The community was split, with D ahead of B (56 to 39), so the vote is directional rather than decisive. AMEJack voted B on the grounds that provisioned concurrency raises cost, which is a fair warning but ignores that provisioned capacity is what removes cold-start latency here. bujuman linked the official provisioned concurrency pricing page, acknowledged the extra charge, and still selected D because both cost and latency had to be satisfied. JackyCCK narrowed the field to B or D after realising that the memory recommendation from AWS Compute Optimizer shortens duration and therefore lowers cost, which is the key insight that makes D coherent.

Official Reference

These links open the official vendor documentation in a new tab.

Exam Strategy

On SAA-C03, treat "maintain latency" or "predictable performance" for Lambda as a provisioned-concurrency trigger, and treat "CPU intensive" as a memory-increase trigger. Reserved concurrency is the distractor because it sounds like reserved capacity but is really a concurrency budget control, not a warm-up mechanism.

Frequently Asked Questions

Why is reserved concurrency (option B) wrong if reserved concurrency is free?

Reserved concurrency only carves a concurrency budget out of the account limit and does not pre-initialize execution environments, so cold-start latency remains and a marketing-event burst can be throttled.

Does increasing Lambda memory really reduce compute cost for CPU-intensive code?

Yes. Lambda allocates vCPU proportionally to memory, so CPU-bound functions finish faster; the shorter GB-second duration often outweighs the higher per-millisecond rate.

More SAA-C03 FAQ →

Community Votes

D
59%
B
41%

59% of anonymous learners picked answer D. Votes are pick records left by other test-takers — they are not the verified answer.

Discussion 16 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 16

  1. JonJon03 👍 8 Answered B
    eserved concurrency — It guarantees the maximum number of concurrent instances for the function which can be invoked. When a function has being with a reserved concurrency configuration then no other lambda function within the same AWS account and region can use that concurrency. There is no charge for configuring reserved concurrency for a function. Provisioned concurrency — This concurrency initializes a requested number of execution environments so that they are prepared to respond immediately to your function’s invocations. Note that configuring provisioned concurrency incurs charges to your AWS account.
  2. alawada 👍 4 Answered D
    Provisioned Concurrency keeps the Lambda functions initialized and ready to process incoming events, reducing the cold start latency associated with spinning up new execution environments.
  3. Sivaeas 👍 3 Answered D
    To reduce compute costs and maintain service latency for customers while using AWS Lambda functions for processing CPU-intensive tasks, you can consider the following strategies: Optimize Lambda Function Configuration: Adjust the memory allocation for Lambda functions to better match the CPU requirements of your workload. Higher memory configurations provide more CPU power. Tune the timeout settings to match the expected processing time of your workload. This prevents unnecessary over-provisioning and reduces costs. Fine-tune the concurrency settings to control the number of concurrent executions based on your workload's characteristics. Use Provisioned Concurrency: AWS Lambda's provisioned concurrency feature allows you to preallocate a number of execution environments to handle incoming requests instantly. This can help reduce cold starts and maintain consistent performance, especially during peak events.
  4. bujuman 👍 3 Answered D
    Also Lambda provisioned concurrency incure additionnal Account charges (https://docs.aws.amazon.com/lambda/latest/dg/provisioned-concurrency.html), it's the best option because it is stated: - The company wants to reduce the compute costs and to maintain service latency for its customers. So maintaining service latency while reducing compute cost is requested. That being said, Lambda optimization is not a trivial task, that's why one should rely on AWS Compute Optimizer recommendations to analyze usage and find the best fit. Please read following for more insights: https://aws.amazon.com/blogs/compute/optimizing-aws-lambda-cost-and-performance-using-aws-compute-optimizer/
  5. LeonSauveterre 👍 3 Answered B
    Let's think - Should we increase or decrease memory? Know that CPU resources scale proportionally with memory allocation in AWS Lambda, so increasing memory will help reduce the execution time for CPU-intensive tasks. Shorter execution times lower the overall compute cost. Then, A & C are out. About option D, it ensures performance and prevents cold starts, yes, but it also incurs additional costs to keep Lambda execution environments pre-warmed, even when not in use. Reserved concurrency is more cost-effective for sustained workloads.
Show 11 more comments
  1. dragossky 👍 2 Answered D
    Reserved concurrency – This represents the maximum number of concurrent instances allocated to your function. When a function has reserved concurrency, no other function can use that concurrency. Reserved concurrency is useful for ensuring that your most critical functions always have enough concurrency to handle incoming requests. Configuring reserved concurrency for a function incurs no additional charges. Provisioned concurrency – This is the number of pre-initialized execution environments allocated to your function. These execution environments are ready to respond immediately to incoming function requests. Provisioned concurrency is useful for reducing cold start latencies for functions. Configuring provisioned concurrency incurs additional charges to your AWS account.
  2. osmk 👍 2 Answered A
    When a large number of messages are in the SQS queue, Lambda scales out, adding additional functions to process the messages. The scale out can consume the concurrency quota in the account. To prevent this from happening, you can set reserved concurrency for individual Lambda functions. This ensures that the specified Lambda function can always scale to that much concurrency, but it also cannot exceed this number. https://docs.aws.amazon.com/lambda/latest/operatorguide/computing-power.html
  3. AMEJack 👍 2 Answered B
    Provioned concurrency will increase cost and we need to maintain latency not to decrease latency. Also, provisioned concurrency cost is very related to memory, so if you use provisioned and increase memory this will affect the cost dramatically.
  4. Salilgen 👍 2 Answered D
    IMO answer is D. The number of messages in message queue is constantly increasing: then, to maintain service latency you need to configure provisioned concurrency. For cpu-intensive job you can optimize the duration increasing memory
  5. JackyCCK 👍 2
    Increase the memory according to AWS Compute Optimizer recommendations --> so we can lower the duration of lambda function to reduce the cost. The ans must be between B & D
  6. JA2018 👍 1 Answered D
    Provisioned concurrency: For scenarios with predictable high traffic like marketing events, using provisioned concurrency ensures Lambda functions are always ready to process requests, maintaining low latency. Increase memory based on Compute Optimizer recommendations: Since the Lambda functions are CPU intensive, increasing allocated memory based on AWS Compute Optimizer suggestions can often improve performance and reduce costs by allowing the functions to process more data without hitting CPU throttling.
  7. tch 👍 1 Answered B
    • Reserved concurrency – This represents the maximum number of concurrent instances allocated to your function. When a function has reserved concurrency, no other function can use that concurrency. Reserved concurrency is useful for ensuring that your most critical functions always have enough concurrency to handle incoming requests. Configuring reserved concurrency for a function incurs no additional charges.
  8. FlyingHawk 👍 1 Answered D
    I vote D instead of B because of constantly increasing number of messages and to maintain service latency for its customer although provisioned concurrency incurs the charges to your AWS account. If the reduce of comput costs is more important than service latency , B is correct.
  9. Gorkha 👍 1 Answered D
    for reserved we don't know the incoming traffic as traffic is increasing so we choose provisioned and as it realted to cpu so compute optimizer.
  10. 1dd 👍 1
    Reserved concurrency its no charges reduce the computation cost, "latency for its customer" then I'll go for A
  11. asdfcdsxdfc 👍 3 Answered D
    D is correct

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Submitted for moderation before publishing. Keep it helpful and respectful.

S3 Multi-Region Access Point and Replication Strategy

S-grade

Securing NLB with Security Groups for Minimal Change

S-grade

Most Cost-Effective 7-Year Retention for EBS Snapshots?

S-grade

AWS Forecast vs SageMaker for Time-Series Prediction

S-grade

Practice All SAA-C03 Questions

Access 120 questions with complete answers and detailed explanations.

View Full SAA-C03 Practice Test →

← Back to SAA-C03 Study Guide