Home / AWS Solutions Architect Associate

AWS · SAA-C03

AWS Solutions Architect Associate Practice Exam (SAA-C03)

Free AWS Solutions Architect Associate practice exam questions from all 4 official domains, with every answer option explained.

Start practicing free Take a mock examSee plans

AWS Certified Solutions Architect - Associate (SAA-C03)

Questions
65
Time
130 min
To pass
720 / 1000

Scored on a 100-1000 scale.

Exam Price
$150

What is on the SAA-C03 exam?

Open a domain for its objectives and their practice questions.

Try free SAA-C03 practice questions

Pick an answer to see why each option is right or wrong. A preview of 6 of the 10 free questions, which cover every domain.

Domain 1: Design Secure Architectures

1.1 Design secure access to AWS resourceseasy

A role used by a nightly batch job received broad permissions during development. The team now wants a policy that contains only the actions the job actually performed over recent weeks. Which tool can generate that policy?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Policy generation reviews the role's activity in CloudTrail and produces a policy template with only the permissions the role used in the chosen time frame, which moves it toward least privilege.

Account Factory provisions new accounts from templates. It does not analyze a role's past actions.

GuardDuty reports suspicious activity; it does not write permission policies from usage.

AWS managed policies are a starting point shared by all customers, so they may not be least privilege for this job.

1.2 Design secure workloads and applicationseasy

A contractor for a United States government agency must ensure that the TLS library protecting its application's AWS KMS calls is FIPS 140 compliant. Which SDK configuration change addresses this?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

FIPS endpoints use a TLS library that complies with FIPS 140, and setting this environment variable tells the AWS SDKs to call them.

General endpoints exist only for a few services and just route to us-east-1; they say nothing about FIPS.

KMS is not among the services with a global endpoint, and a global endpoint would not imply FIPS compliance.

Dual stack endpoints are real, but they exist so clients can use IPv4 or IPv6, not to provide a FIPS 140 TLS library.

1.3 Determine appropriate data security controlseasy

Users of a shared S3 bucket sometimes delete or overwrite reports by mistake. The team wants to be able to bring back an earlier copy of any affected object. What should it enable?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Encryption protects data at rest but does nothing to keep deleted or overwritten content.

With versioning, a delete adds a delete marker instead of removing the object, and an overwrite creates a new version, so earlier versions can be restored.

Block Public Access stops public exposure; it keeps no earlier copies of objects.

Expiration deletes objects on a schedule, which works against recovering them.

Domain 2: Design Resilient Architectures

2.1 Design scalable and loosely coupled architectureseasy

A broadcaster's migration team will move several hundred Java services off on-premises virtual machines. During the move, each service is packaged into a container image and run on Amazon ECS, and its source code is left untouched. Which of the 7 Rs describes this plan?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Repurchase also ends with something new running, but the new thing is a different version or product that replaces the application. The broadcaster keeps its own Java services and changes only how they run.

Retire is for applications you want to decommission or archive, so their servers can be shut down. The team is keeping every service in use and moving it, so nothing is being retired.

Replatform moves an application to the cloud while introducing some level of optimization. Migrating virtual machines into containers without making any code changes is one of its listed use cases, which is what the team is doing.

Retain fits applications that stay in the source environment or are not ready to migrate. These services are leaving the on-premises virtual machines for Amazon ECS, so they are not being kept where they are.

2.1 Design scalable and loosely coupled architectureseasy

A retailer runs separate services for /cart and /search and wants one public endpoint that sends each request to a different target group based on its URL path. Which load balancer fits?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

An ALB works at layer 7 and evaluates listener rules against the request content, so a path condition can forward /cart and /search to different target groups.

Listeners are defined by protocol and port, not URL path, and Classic Load Balancers are the previous generation that AWS recommends migrating away from.

Gateway Load Balancers exist to insert virtual appliances such as firewalls into the traffic path; they do not do HTTP path routing for application services.

An NLB is the right choice for layer 4 TCP/UDP traffic at very high rates, but at layer 4 it never looks at URL paths, so it cannot route on /cart versus /search.

2.2 Design highly available and/or fault-tolerant architectureseasy

An operations team patches its web servers by logging in over SSH and running updates on each running instance. Over time the servers have drifted apart in configuration. Which change follows an immutable infrastructure model?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Running the script everywhere at once makes changes more uniform, but it is still an in-place modification of running servers, the exact anti-pattern immutable infrastructure rules out.

Adding capacity changes how many servers exist, not how they are changed. The new servers would be patched over SSH like the rest, so nothing about the model becomes immutable.

Immutable infrastructure forbids in-place updates, patches and configuration changes on production resources. A new, updated set of resources is deployed and traffic moves to it, which also removes the drift because every server comes from the same known image.

Snapshots give a point-in-time copy to fall back on, which reduces risk, but the servers are still changed in place and keep drifting. It protects against a bad patch rather than removing the practice causing drift.

The other 4 free questions, by domain:

Domain 3: Design High-Performing Architectures · 2 questions
3.2 Design high-performing and elastic compute solutionseasy

An insurer is moving its claims system to AWS. The system's components exchange messages through an on-premises Apache ActiveMQ broker using JMS, and the team wants to migrate without rewriting that messaging code. Which service fits?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Amazon MQ is the managed broker AWS recommends for migrating applications that depend on existing broker APIs such as JMS and on engines like ActiveMQ, so the components can keep their messaging code.

SNS is a managed publish-subscribe service, so it fits the idea of messaging. It delivers through topics to endpoints such as queues, functions and email, not as a JMS broker, so the existing code would still have to change.

SQS is AWS's queue for decoupling components and is recommended for new applications, so it is a natural first thought. Using it means moving the components onto its own web services API, which is the rewrite the team wants to avoid.

AWS Batch also manages work queues, which can make it sound related. It provisions compute for submitted batch jobs and does not provide a message broker for application components.

3.1 Determine high-performing and/or scalable storage solutionseasy

A firm is moving a Windows-based document management application to EC2. The application keeps its files on SMB shares and depends on native Windows file system features. The team wants a fully managed file service. Which should it choose?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

FSx for Lustre is a managed high-performance file system, but it is built for Linux clients and speed-critical workloads such as HPC, not Windows SMB shares.

Transfer Acceleration speeds up long-distance uploads to an S3 bucket; S3 is object storage and does not provide Windows file shares for the application.

EFS is the usual answer for managed shared file storage, but it serves NFS clients and is not supported on Windows-based EC2 instances.

FSx for Windows File Server provides fully managed file servers built on a native Windows file system with SMB support, which is what a Windows application that relies on Windows file features needs.

Domain 4: Design Cost-Optimized Architectures · 2 questions
4.2 Design cost-optimized compute solutionseasy

A cost review suggests that many EC2 instances in a production account are larger than they need to be. The team wants recommendations for better instance types and sizes based on each instance's actual CPU, memory, network and disk utilization, with projected utilization for each recommendation. Which service should it use?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Budgets tracks spend and usage against thresholds and can alert or act, but it does not analyze instance utilization or recommend instance types.

CUR delivers detailed billing line items to S3 for analysis, which helps find expensive instances. It contains no rightsizing recommendations or projected utilization.

Compute Optimizer analyzes resource configuration and CloudWatch utilization metrics to produce rightsizing recommendations and flag idle resources. It shows projected utilization for each option so the team can weigh price against performance, after opting in.

Cost Explorer is the natural place to see that EC2 spend is high, and it recommends Reserved Instance purchases. It analyzes costs and usage rather than each instance's utilization metrics, so it does not say which smaller type or size fits.

4.1 Design cost-optimized storage solutionseasy

A finance team applied a project tag to every S3 bucket and EBS volume in a standalone AWS account. A week later, Cost Explorer still offers no way to filter or group storage costs by that tag. What is the missing step?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Resource tags do not show up in billing tools on their own. A user-defined tag has to be activated as a cost allocation tag before Cost Explorer and the cost allocation report can use it, and it can take up to 24 hours to appear.

Storage Class Analysis can group objects by tag, which makes it sound relevant. It studies access patterns to guide lifecycle transitions; it does not feed tags into Cost Explorer, and it would not cover EBS volumes at all.

Consolidated billing combines bills and discounts across many accounts, which is a cost feature but solves a different problem. A standalone account can already activate cost allocation tags without joining an organization.

Budgets are useful for alerting on project spend, so they seem related. A budget does not make an inactive tag available as a dimension in Cost Explorer; the tag still has to be activated first.

Practice all 10 free questionsGet 150+ questions

SAA-C03 question bank coverage

DomainExam weightFree sampleMaster (full bank)
1. Design Secure Architectures30%358
2. Design Resilient Architectures26%351
3. Design High-Performing Architectures24%247
4. Design Cost-Optimized Architectures20%239
Total100%10195

Looking for a SAA-C03 practice exam? Start with the free questions above, then take the timed mock exam to rehearse the SAA-C03 format under exam conditions.

How these SAA-C03 questions are made

The domains, weights and objectives come from AWS's official SAA-C03 exam guide (v1.1). Questions are original and written only from the vendor's official prep material, and each one cites the passage it is based on. Every question is then answered independently, without seeing its key, and every answer option's explanation is checked for accuracy before it is published.