Your architecture practice round

{{ summaryValue }}

{{ summaryLine }}

Build an SAA-C03 architecture round

Unofficial AWS Certified Solutions Architect – Associate (SAA-C03) practice · SAA-C03 HTML guide verified September 12, 2026. Independently authored; not affiliated with or endorsed by AWS.

32 original English scenarios practise architecture trade-offs across all four domains. This is untimed single-answer practice, sampled uniformly without replacement. The real exam has 65 questions in 130 minutes, including multiple-choice and multiple-response questions. This bank does not reproduce full-exam breadth or prove hands-on competence.

Official guide and format · Official preparation resources

Blueprint weights: security 30%; resilience 26%; performance 24%; cost 20%. This curated bank has eight scenarios per domain. Weights are study context, not sampling weights or score multipliers. Difficulty is uncalibrated; there is no timed or weighted mode. AWS recommends at least one year designing AWS cloud solutions; no prior certification is required.

Optional. Blank uses a stable default. The first 128 UTF-16 units of the trimmed seed are used, without splitting code points. The active round keeps its own seed for replay.

{{ examSetupValidation }}

{{ answeredCount }} of {{ questionCount }} answered · {{ Math.round(progressPercent) }}%

{{ progressStatusLine }}

Question {{ examCursor + 1 }} of {{ questionCount }}

{{ currentQuestion.stem }}

{{ currentQuestion.accessible_clue }}

Answer recorded. Corrections appear when you view results.

Review this practice round

{{ resultLead }}

SAA-C03 · guide verified 2026-09-12 · original bank 2026.09.12. This score describes only the items answered. It is not an AWS scaled score, a pass prediction, or evidence of certification. AWS and its service names are trademarks of Amazon.com, Inc. or its affiliates.

This score covers only this curated round. Uniform sampling and uncalibrated difficulty do not reproduce the full blueprint. The real exam includes multiple-response questions; this untimed single-answer practice does not test those or demonstrate hands-on architecture competence. It cannot predict passing.

{{ readinessReviewTitle }}

{{ readinessReviewLine }}

{{ card.label }}
{{ card.value }}
{{ card.note }}

Missed answer review

  1. {{ item.prompt }}

    Your answer: {{ item.selected_answer_label }}

    Correct answer: {{ item.correct_answer_label }}

    {{ item.explanation }}

    {{ item.source.label.split(' — ')[0] }} · {{ item.source.version }}

Domain performance

Correct and missed answers by official domain for this round. Domains absent from the chart were not sampled. Use the missed-item sources to review that domain, then practise a focused set and check the full official guide.

Loading domain-performance chart…

{{ examChartStatus }}

{{ examChartExportStatus }}

Choose your next design domain

  • {{ domain.label }}

    {{ domain.weight }}% of official scored content · {{ domain.answerLine }}

    {{ domain.study }}

Answer ledger

Question ID Prompt Topic Your answer Correct answer Outcome Explanation Source Attempt identity Copy
{{ row.question_id }} {{ row.prompt }} {{ row.category_label }} ({{ row.category_id }}) {{ row.selected_answer_label }} ({{ row.selected_answer_id }}) {{ row.correct_answer_label }} ({{ row.correct_answer_id }}) {{ row.outcome_label }} {{ row.explanation }} {{ row.source ? `${row.source.label.split(' — ')[0]} · ${row.source.version}` : 'Official guide' }} {{ sourceUrl(row.question_id) }} {{ attemptIdentity }}

{{ examLedgerExportStatus }}

AWS Cloud Practitioner Exam Prep Practise AWS Cloud Practitioner CLF-C02 concepts with original questions, immediate explanations and focused reviews of missed answers.
Cloud Egress Cost Calculator Estimate cloud egress spend from measured traffic and a marginal rate card with verified reductions, peak-month stress testing, and budget variance.
Cloud NAT Gateway Cost Calculator Estimate monthly cloud NAT cost from gateway time and routed traffic while comparing offload savings, route charges, and budget impact.
Object Storage Cost Calculator Estimate monthly object storage cost from capacity and request volume with retrieval, egress, lifecycle and editable provider rate assumptions.

A standby database and a read replica can solve different problems even though both hold another copy of data. One may support failover, while the other takes reporting traffic away from the primary database. Choosing services on Amazon Web Services (AWS) means matching those behaviors to the workload rather than treating all replication as interchangeable.

The AWS Certified Solutions Architect – Associate exam, SAA-C03, focuses on designing and improving architectures that are secure, resilient, high-performing, and cost-optimized. A suitable answer must meet the requirements together. Adding capacity does not repair an excessive permission, and reducing a storage bill does not help when the chosen storage cannot meet the required access time.

Three distinctions make architecture scenarios easier to reason about.

Availability and recovery
Availability concerns whether a service can keep serving requests. Recovery concerns restoring service or data after disruption. A failover arrangement and a backup policy address different needs.
Read consistency
A read may need to reflect the latest successful write, or the application may tolerate older data while changes propagate. That requirement affects which database access paths are suitable.
Least privilege
Access should allow only the actions and resources needed for the task. A credential that works is not necessarily an appropriately restricted credential.

Failure scope matters as much as the number of resources. Availability Zones (AZs) are isolated locations within an AWS Region. Spreading an application across AZs addresses a different disruption from losing the Region itself. A regional recovery plan must account for the application, its data, and the capacity needed after traffic moves. Keeping a reduced-capacity but functioning recovery environment running can reduce the work required during recovery, but recovery objectives still need testing.

Application behavior also determines whether a service fits. Amazon Simple Queue Service (SQS) standard queues can separate incoming jobs from slower processing, but delivery is at least once and messages can arrive out of order. A payment-record worker therefore needs idempotent processing, meaning repeated delivery does not create a second business transaction. Choosing a durable queue does not remove that application responsibility.

Cost questions require the same attention to constraints. A cheaper storage option may require a restore before data can be read. Extra recovery capacity can increase ongoing costs. The useful comparison includes the access pattern, acceptable recovery work, and resources that remain billable, rather than assuming that the lowest individual price produces the lowest acceptable total cost.

AWS recommends at least one year of hands-on experience designing cloud solutions for this certification. Scenario practice is useful for learning which requirement rules an alternative out, then checking that reasoning against documentation and practical work. Practice scores are educational feedback, not certification, a passing forecast, or evidence of hands-on architecture competence.

How to Use This Tool:

Choose a mixed architecture round or concentrate on one design domain. The bank contains 32 independently authored English scenarios, each with one answer to choose from four alternatives. This is unofficial, untimed practice, not affiliated with or endorsed by AWS.

  1. Choose a Domain / practice focus and Question count. The initial selection is eight questions from All domains · 32 scenarios. Mixed rounds offer 2, 4, 8, 16, or 32 questions; a single domain offers 2, 4, or 8. Use a count available for the selected focus before starting.
  2. Select Feedback timing. After each answer reveals the correction, explanation, and supporting source before you continue. At the end records each answer without revealing its correctness until the completed review.
  3. Enter an optional Replay seed under Advanced when you need a reproducible round. Leaving it blank uses a stable default rather than choosing a fresh random seed each time.
  4. Select Start exam practice, read the scenario's constraints, and choose one answer. Selecting an option immediately records and locks it in either feedback mode. Continue with Next question, then select View results after answering every question.
    End round clears the active answers and returns to setup. It does not produce a partial final review.
  5. Review the explanations in Missed answer review and the relevant source documents. Use Retry missed to practise those questions again, or choose Practise domain beside a domain to return to setup with that domain and four questions selected.

Interpreting Results:

During a round, the main figure counts questions answered, not answers correct. In the completed review, it changes to correct answers out of all questions in the round. Perfect round means no misses in that selection; it is not a readiness grade.

Domain performance shows correct and missed answer counts for the sampled domains. Compare each count with the number answered in that domain rather than reading a longer bar as greater ability. A domain marked Not sampled in this round has not been assessed. The Answer ledger supplies the selected answer, correct answer, explanation, and source for checking individual decisions, including correct guesses.

Replay same seed repeats the active round's questions and option order. Try a new seed draws from the same limited bank and can include familiar questions again. A full 32-question round already contains the whole bank, so another full round cannot introduce unseen scenarios. Higher replay scores may reflect familiarity as well as improved understanding.

Use the missed explanations to identify a specific gap, such as confusing database failover with read scaling, rather than treating a domain total as a complete skill profile. If the chart is unavailable, the written domain counts and answer ledger remain useful; Retry chart attempts to load the visualization again.

Technical Details:

Each scenario has one authored best answer for its stated requirements. Every practice question has the same scoring value, so the score measures the selected answers against that key. It does not measure the quality of an actual deployment or reproduce official scaled scoring.

Correct answers and round progress

Let N be all questions in the round, A the number answered, and C the number answered correctly. Each correct answer earns one point and each incorrect answer earns zero, with no deductions or partial credit. The possible score is zero to N points. The practice accuracy percentage divides correct answers by all round questions.

P = 100 × C N

Progress measures completion instead of correctness. Its percentage uses the number answered, while the number missed counts only submitted incorrect answers.

G = 100 × AN M = A − C

P is practice accuracy, G is progress, and M is the missed-answer count. Both percentages range from 0% to 100%. Unanswered questions remain in N but do not increase M. Accuracy keeps the unrounded ratio, and the completed headline preserves the exact correct/total counts. The visible progress percentage rounds to the nearest whole percent, with an exact half-percent tie rounding upward. Completion depends on every question being answered, not a rounded percentage.

For example, after three answers in an eight-question round, with two correct and one incorrect, progress is 37.5%, shown as 38%. Practice accuracy is 25% because the denominator is still eight. There is one miss and five unanswered questions; the final review is not yet available.

Condition Review behavior Meaning
A < N No completed review At least one question remains unanswered.
A = N and C < N Start with the missed answers At least one submitted answer was incorrect.
A = N and C = N Perfect round Every answer in this round was correct.

These are completion and review conditions, not pass bands. Domain counts use the same one-point answer decisions, without additional domain multipliers.

Official blueprint and practice coverage

The four official domain weights describe the distribution of scored SAA-C03 content. The curated bank instead contains eight scenarios in each domain. The topic examples below describe the practice material, not an exhaustive exam syllabus.

Official domain Official weight Practice scenarios Selected practice topics
Design Secure Architectures 30% 8 Temporary credentials, federation, private access, key permissions, and credential rotation.
Design Resilient Architectures 26% 8 Failover, queues, duplicate handling, failed-job retries, versioning, and regional recovery.
Design High-Performing Architectures 24% 8 Shared files, read scaling, indexes, consistency, compute placement, networking, and ingestion.
Design Cost-Optimized Architectures 20% 8 Archive access, storage tiering, interruptible compute, commitments, traffic paths, and capacity billing.

The official exam has 65 questions in 130 minutes, including single-answer multiple-choice questions and multiple-response questions. Of these, 50 affect the score and 15 are unidentified unscored questions. AWS reports a scaled score from 100 to 1,000, with a minimum passing score of 720. That is not a 72% raw-score rule, and there is no supported conversion from this practice percentage to an AWS scaled score.

Unlike the practice round's treatment of unanswered items while work is in progress, the official exam scores unanswered questions as incorrect. AWS evaluates passing overall rather than requiring a separate pass in every domain. The untimed, single-answer practice bank has no official pass threshold, multiple-response mode, or calibrated difficulty.

The bank records a September 12, 2026 reference check for the SAA-C03 guide and its technical sources. That is a reference snapshot date, not an exam launch date. Eight scenarios per domain do not establish complete coverage of the domain's tasks or services, and the questions do not update themselves when AWS documentation changes.

Sampling and reproducibility

Selection is uniform without replacement within the chosen question pool. A question does not repeat within one round, and there are no official domain quotas. Questions are fixed when the round starts rather than chosen adaptively as you answer. Selecting all 32 questions gives each domain 25% of the questions, rather than the official 30%, 26%, 24%, and 20% distribution.

Keeping the same focus, count, feedback timing and seed repeats the questions and their order while the bank and saved-attempt format remain unchanged. Changing feedback timing can change the round even with the same seed. Surrounding spaces are ignored, and long seed text is shortened at a whole character. The allowance holds 128 basic letters or digits; many emoji take two places.

A missed-only retry has its own denominator, equal to the number of missed questions. This can produce a one-question retry even though the ordinary setup menus start at two questions. The retry measures that restricted selection, not the original full round.

Architecture constraints behind the answers

For an application on Amazon Elastic Compute Cloud (EC2) that needs access to a private Amazon Simple Storage Service (S3) bucket, temporary credentials and restricted permissions solve separate parts of the requirement. An AWS Identity and Access Management (IAM) role attached through an instance profile avoids distributing a long-lived user key. Its permissions still need to restrict the allowed actions and resources; replacing the credential mechanism alone does not establish least privilege.

For DynamoDB, an alternate query path and a current-data requirement are different considerations. A global secondary index (GSI) supports eventually consistent reads, not strongly consistent reads. When the application knows the base-table key and needs a read that reflects successful prior writes, a strongly consistent base-table read fits that requirement. Adding a filter or requesting a different consistency setting cannot make a GSI strongly consistent.

Network scope also changes an answer. An S3 gateway endpoint gives clients in a virtual private cloud (VPC) a same-Region S3 route without a network address translation (NAT) gateway or internet gateway. It is not a general solution for on-premises access through that VPC. Route tables and applicable identity, bucket, and endpoint permissions still matter. Avoiding NAT traversal removes NAT processing charges for that eligible path, not every S3 or network charge.

For internet-bound traffic crossing AZs to a NAT gateway, same-AZ routing can remove the cross-AZ part of the path. An additional gateway also brings its own hourly charges, and NAT data processing remains billable. The total-cost choice depends on traffic and availability requirements; no universal saving or break-even traffic level follows from choosing more gateways. Service availability, supported configurations, quotas, and current pricing should be checked for the intended Region before applying a practice answer to a real workload.

Privacy Notes:

Answers are scored in the browser, but the saved attempt is not private storage. The page URL carries the seed, round information, and responses in an encoded, unencrypted form. A recipient of that URL can recover the attempt, and URLs can remain in browser history or reach the site when opened or reloaded.

Use a non-sensitive seed, not a password, AWS access key, or personal identifier. Exported answer ledgers also contain responses and attempt information. Charts and document exports can require network-loaded resources, and opening a reference contacts its publisher. Fully offline or zero-network use is not guaranteed.

A malformed saved URL or an incompatible question-bank or round version can prevent restoration. When Saved attempt unavailable appears, start a new round from setup. If a warning says the current attempt could not be saved in the URL, keep the page open to continue; do not rely on reloading that URL to recover the latest answers.

Worked Examples:

Six correct answers, then a targeted retry

Complete an eight-question mixed round with six correct answers and two incorrect answers. The completed summary is 6/8, equivalent to 75% practice accuracy, with two questions in the missed review. Choosing Retry missed starts a two-question round. Answering both correctly produces 2/2 for that retry, not a revised 8/8 for the original selection. Read each explanation to check the changed reasoning before using another mixed round to practise beyond those two familiar questions.

Offloading reports without changing writes

In the reporting scenario, an Amazon Relational Database Service (RDS) MySQL primary is overloaded by analytics reads, and the reports can tolerate replication lag. Choosing a read replica and routing those reports to its endpoint is the authored correct answer. Routing reports to the standby of a Multi-AZ DB instance deployment is marked incorrect because that single standby does not serve read traffic. This restriction does not apply to every deployment called Multi-AZ; a readable Multi-AZ DB cluster is a different configuration. The explanation directs attention to read capacity, replication lag, and the exact deployment type rather than treating every additional database copy as equivalent.

References: