CoStudy

HomeCertificationsAWS Certified Developer DVA-C02 › Troubleshooting and Optimization

Troubleshooting and Optimization — AWS Certified Developer DVA-C02 practice questions

33 multiple-choice questions and 12 flashcards on Troubleshooting and Optimization, about 22% of the AWS Certified Developer DVA-C02 bank. Every one carries a written rationale.

Written and maintained by Nick Burton · last updated 2026-08-22 · how we write and review questions

What this chapter covers

Troubleshooting and Optimization is one of 4 chapters in CoStudy's AWS Certified Developer — Associate (DVA-C02) bank, and it holds 33 of the bank's 150 multiple-choice questions — roughly 22% of the total. That proportion is not arbitrary: chapters follow the certifying body's published exam outline, and the number of questions in each is set by that domain's published weight, so the share of your practice time this chapter takes matches the share of the real exam it accounts for.

Studying by chapter is worth doing once you have a diagnostic score. A single overall percentage tells you whether you are close; it does not tell you which domain is dragging. Working a weak chapter in isolation, and re-testing it in isolation, is the fastest way to move a score that has stalled — and it is why the mock exams in CoStudy report by domain rather than as one number.

Free Troubleshooting and Optimization practice questions

10 questions drawn from this chapter, with the full rationale shown — the controlling principle behind the right answer, and why each wrong option tempts and fails.

Lambda is 1 GB memory and CPU-bound at 800 ms. Doubling memory typically does what?

  1. Has no effect on CPU allocation
  2. Cuts CPU in half by design here
  3. Doubles CPU and often lowers cost
  4. Guarantees doubled invocation cost

Answer: C — Doubles CPU and often lowers cost

A) Memory and CPU are linked in Lambda. C) Correct — proportional CPU allocation often shortens runtime enough that total cost drops. B) Wrong direction entirely. D) Higher rate times shorter duration can net cheaper.

You need an alarm when Lambda error rate exceeds 1% over 5 minutes. BEST expression?

  1. Static threshold on Invocations only
  2. Metric math: Errors/Invocations*100
  3. Static threshold on Errors at 1
  4. Composite alarm on other alarms

Answer: B — Metric math: Errors/Invocations*100

A) Invocations doesn't indicate errors. C) A single error may be normal at scale. B) Correct — metric math derives an error rate independent of traffic volume. D) Composite alarms combine, not compute.

Which SQS setting reduces empty-receive polling costs to a Lambda-based consumer?

  1. Delay seconds of 900 seconds
  2. Short polling with 0-second wait
  3. Message retention of 14 days
  4. Long polling with 20-second wait

Answer: D — Long polling with 20-second wait

D) Correct — long polling waits up to 20s for a message and cuts empty responses. B) Short polling returns immediately and multiplies request cost. C) Retention doesn't affect polling frequency. A) Delay hides new messages but doesn't lower poll cost.

Which CloudWatch alarm math evaluates a Lambda error RATE rather than raw count?

  1. Threshold on Invocations only
  2. Threshold on Errors only at 1 event
  3. Composite alarm over all Lambdas
  4. Metric math: Errors / Invocations * 100

Answer: D — Metric math: Errors / Invocations * 100

A) Invocations is a volume metric, not an error signal. B) A single error may be normal at high volume. D) Correct — metric math derives a rate that scales with traffic. C) Composite alarms combine other alarms.

Which Lambda concurrency setting pre-initializes containers so cold-start latency is avoided?

  1. Reserved concurrency
  2. Unreserved concurrency
  3. Provisioned concurrency
  4. Account concurrency limit

Answer: C — Provisioned concurrency

A) Reserved caps the max concurrent executions but does not pre-warm. C) Correct — provisioned concurrency keeps N initialized containers ready to serve. B) Unreserved is the leftover pool for non-configured functions. D) The account limit is the ceiling, not a warming feature.

Which retry pattern does the AWS SDK apply by default when it receives a throttling response?

  1. Immediate constant-interval retry
  2. Exponential backoff with jitter
  3. Linear backoff with fixed step
  4. No retry — surfaces the error

Answer: B — Exponential backoff with jitter

A) Constant retries stampede the throttled service. B) Correct — SDK default is exponential backoff with jitter to spread retries. C) Linear backoff is not the SDK default. D) The SDK does retry retryable errors before raising.

SQS visibility timeout is 30s but processing takes 45s. Consequence?

  1. Message reappears at 30s to peer
  2. Message stays hidden until deleted
  3. Message auto-deleted at 30s mark
  4. Message immediately moves to DLQ

Answer: A — Message reappears at 30s to peer

B) Hidden means invisible; not indefinitely so. A) Correct — visibility expires and the message becomes visible again, risking duplicate processing. C) Not auto-deleted on timeout. D) DLQ requires exceeding maxReceiveCount.

Which retention default applies to a NEW CloudWatch log group if you set nothing?

  1. 1 day retention shortest limit
  2. 30 day retention recommended
  3. 90 day retention target period
  4. Never expire — retain forever

Answer: D — Never expire — retain forever

A) 1 day is the shortest configurable value, not the default. B) 30 days is a common recommendation, not the default. C) 90 days is also configurable. D) Correct — new log groups default to Never Expire; this drives cost surprises.

DynamoDB scans are slow and expensive. Data is filtered by a non-key attribute. BEST fix?

  1. Create a GSI on that attribute
  2. Increase table RCUs to speed scan
  3. Run scans only during off-hours
  4. Enable parallel scan on the table

Answer: A — Create a GSI on that attribute

B) More RCUs only make expensive scans faster, not cheaper. A) Correct — Query on a GSI is orders of magnitude cheaper than Scan. C) Off-hours doesn't help user latency. D) Parallel scan is still expensive.

Which HTTP status code from an SDK call indicates client-side throttling by the target AWS API?

  1. HTTP 429 Too Many Requests
  2. HTTP 403 Forbidden
  3. HTTP 401 Unauthorized
  4. HTTP 500 Internal Server

Answer: A — HTTP 429 Too Many Requests

C) 401 is unauthenticated. B) 403 is access denied. A) Correct — 429 TooManyRequests signals throttling; SDKs retry with backoff. D) 500 signals a server-side error.

Troubleshooting and Optimization flashcards

3 cards from the 12 in this chapter.

CloudWatch Logs?

Centralized log management. Lambda, EC2, ECS, etc. send logs.

Lambda timeout error?

Function exceeded configured timeout (max 15 min). Increase timeout or optimize code.

How to reduce Lambda cold starts?

Use Provisioned Concurrency, smaller deployment package, lighter runtime, init code outside handler.

Practise the full chapter

These are a sample. The full Troubleshooting and Optimization chapter runs 45 items with per-chapter progress tracking, on the web and in the iOS app.

Open AWS Certified Developer DVA-C02 in CoStudy →

Other AWS Certified Developer DVA-C02 chapters

All AWS Certified Developer DVA-C02 practice questions →