SAA-C03 Exam Questions

Free AWS Solutions Architect Exam Questions and Answers

Updated October 13, 2026  |  20 questions with full answers

The short answer

SAA-C03 is 65 questions in 130 minutes with a 720/1000 passing score. It is a scenario exam: almost every question describes an architecture and asks you to choose the option that best meets a stated requirement. The questions below come from our 975-question SAA-C03 bank, covering all four domains, with the answer and explanation.

SAA-C03 exam facts

Questions65 (multiple choice and multiple response)
Time limit130 minutes
Passing score720 / 1000
Cost$150 USD
DeliveryPearson VUE test centre or online proctored
PrerequisitesNone formally, but AWS recommends hands-on experience

What the exam covers, domain by domain

Domain Weight What it tests
Domain 1 - Design Secure Architectures30%IAM, encryption, network security, data protection, key management.
Domain 2 - Design Resilient Architectures26%High availability, multi-AZ and multi-region, decoupling, disaster recovery.
Domain 3 - Design High-Performing Architectures24%Compute, storage, database and networking choices for performance.
Domain 4 - Design Cost-Optimized Architectures20%Storage classes, instance purchasing options, cost-aware architecture.

20 free SAA-C03 exam questions with answers

Real questions from our 975-question SAA-C03 bank, spread across every domain. The correct answer is marked and every question carries an explanation, so you can check the reasoning and not just the result. The interactive mocks serve the full bank, timed and scaled-scored.

Question 1 Design Resilient Architectures Choose 2

A solutions architect is designing a fault-tolerant system for processing IoT device data. The system must ensure no data loss if a processing component fails and must be able to reprocess data from a specific point in time. Which TWO AWS services should be used as part of this design? (Select TWO)

  • ✓ Amazon Kinesis Data Streams
  • ✓ Amazon Simple Queue Service (SQS) Standard queue
  • ○ AWS Lambda with on-demand concurrency
  • ○ Amazon Simple Notification Service (SNS)
  • ○ Amazon Kinesis Data Firehose
Answer: A. Amazon Kinesis Data Streams | B. Amazon Simple Queue Service (SQS) Standard queue
Amazon Kinesis Data Streams (A) provides durable, ordered data streams with a retention period (24 hours by default, extendable to 365 days), allowing replay from a past point for reprocessing, which is key for fault tolerance. Amazon SQS (B) provides at-least-once delivery and message durability, preventing data loss if a processor fails. SNS (D) is a pub/sub service but lacks built-in retention for replay. Kinesis Data Firehose (E) is for loading to destinations, not for replayable processing. Lambda (C) is a compute option but doesn't provide the data durability layer.
Question 2 Domain 1: Design Secure Architectures

A company wants to grant temporary access to an AWS S3 bucket for an external contractor. The access should be limited to specific operations and expire after 3 days. Which approach is MOST secure?

  • ○ Create an IAM user for the contractor and attach a policy with the necessary permissions.
  • ✓ Use IAM roles and assume role with temporary security credentials.
  • ○ Share the access keys of an existing IAM user with the contractor.
  • ○ Create a bucket policy that allows access from the contractor's IP address.
Answer: B. Use IAM roles and assume role with temporary security credentials.
IAM roles with temporary security credentials are the most secure for granting temporary access because credentials expire automatically and do not require managing long-term access keys. Option A involves long-term credentials which are less secure for temporary needs. Option C is insecure as it shares permanent credentials. Option D is not user-specific and relies on IP which can be spoofed or change.
Question 3 Domain 2: Design Resilient Architectures

An e-commerce company uses an Amazon RDS for MySQL database for its application. The database is currently in a single Availability Zone. To improve availability and ensure automated failover in case of an AZ outage, the Solutions Architect must implement a solution with minimal downtime and no application changes. What should the architect do?

  • ○ Create a read replica in another AZ and configure the application to use the replica if the primary fails.
  • ✓ Modify the RDS instance to use a Multi-AZ deployment.
  • ○ Create a database snapshot and restore it in another AZ, then update the application endpoint.
  • ○ Migrate the database to Amazon Aurora with a replica in another AZ.
Answer: B. Modify the RDS instance to use a Multi-AZ deployment.
Modifying an existing RDS instance to use a Multi-AZ deployment is the simplest way to achieve high availability with automated failover and minimal downtime. RDS handles the creation of a synchronous standby replica in another AZ and manages the failover process automatically. The application connection string remains the same, so no application changes are required. Creating a read replica requires manual promotion and application changes. Restoring a snapshot is a manual recovery process, not automated failover. Migrating to Aurora is a larger architectural change, not the minimal solution requested.
Question 4 Domain 3: Design High-Performing Architectures

An e-commerce application uses an Amazon RDS for PostgreSQL database. During sales events, the database experiences high read traffic, causing increased latency for customers. Which solution will improve read performance WITHOUT adding management overhead?

  • ○ Create a read replica and direct read queries to it
  • ○ Upgrade the RDS instance to a larger instance class
  • ○ Implement Amazon ElastiCache for Redis to cache query results
  • ✓ Use Amazon Aurora PostgreSQL with Aurora Replicas
Answer: D. Use Amazon Aurora PostgreSQL with Aurora Replicas
Amazon Aurora PostgreSQL with Aurora Replicas provides up to 15 low-latency read replicas that share the same underlying storage as the primary, offloading read traffic with minimal performance impact and management overhead. A read replica for standard RDS (A) requires separate storage and replication lag management. Upgrading the instance (B) is vertical scaling and does not specifically address read scaling. ElastiCache (C) adds caching but requires application changes and cache management.
Question 5 Domain 4: Design Cost-Optimized Architectures

A company uses AWS Organizations with multiple accounts. They want to centralize cost management and optimize spending. What should they enable?

  • ✓ AWS Cost Explorer with linked accounts
  • ○ AWS Config aggregator
  • ○ AWS Service Catalog
  • ○ AWS Control Tower
Answer: A. AWS Cost Explorer with linked accounts
Option A is correct because Cost Explorer with linked accounts provides consolidated cost and usage data across all accounts, enabling better optimization. Option B is incorrect as Config aggregator is for compliance, not cost management. Option C is incorrect because Service Catalog is for IT service management. Option D is incorrect as Control Tower is for multi-account governance, not specifically cost optimization.
Question 6 Domain 1: Design Secure Architectures

Data uploaded to an S3 bucket must be encrypted in transit. How can this be enforced?

  • ○ Use S3 default encryption with SSE-S3.
  • ✓ Configure a bucket policy to deny HTTP requests.
  • ○ Enable S3 versioning.
  • ○ Use IAM policies to require SSL/TLS.
Answer: B. Configure a bucket policy to deny HTTP requests.
S3 bucket policies can include conditions to deny non-HTTPS requests, enforcing encryption in transit. Option A encrypts data at rest. Option C is for data recovery. Option D is not standard; IAM policies don't directly enforce S3 transport encryption.
Question 7 Domain 2: Design Resilient Architectures

A legacy application running on-premises will be migrated to AWS. It requires a shared file system that is accessible from multiple Amazon EC2 instances in a VPC. The solution must be highly available, durable, and managed. Which AWS service should be used?

  • ○ Amazon S3 with a file gateway interface.
  • ✓ Amazon EFS (Elastic File System)
  • ○ Amazon EC2 instance store aggregated with RAID.
  • ○ An Amazon EBS volume with Multi-Attach enabled.
Answer: B. Amazon EFS (Elastic File System)
Amazon EFS is a fully managed, highly available, and durable network file system (NFS) that can be mounted concurrently by thousands of EC2 instances across multiple AZs within a region. It scales automatically and is designed for high availability. S3 (A) is object storage, not a low-latency shared file system. Instance store (C) is ephemeral and not managed or durable. EBS Multi-Attach (D) is limited to instances in the same AZ and specific Nitro-based instance families.
Question 8 Domain 3: Design High-Performing Architectures Choose 2

Which TWO are characteristics of the Amazon S3 Glacier storage classes (such as Glacier Instant Retrieval) that are relevant for archive performance? (Select TWO.)

  • ○ Millisecond retrieval latency
  • ○ Optimized for frequent access
  • ✓ Lower cost than S3 Standard
  • ✓ Designed for long-term archive
  • ○ Provides the highest durability
Answer: C. Lower cost than S3 Standard | D. Designed for long-term archive
S3 Glacier storage classes (including Instant Retrieval, Flexible Retrieval, Deep Archive) are designed for long-term archive and backup (D) at costs significantly lower than the S3 Standard storage class (C). They do not offer millisecond retrieval; retrieval times range from milliseconds (Instant Retrieval) to hours. They are not optimized for frequent access (B). All S3 storage classes provide the same high durability (11 9's) (E), so this is not a differentiating characteristic for Glacier.
Question 9 Domain 4: Design Cost-Optimized Architectures

A company uses Amazon DynamoDB for a mobile gaming application. Traffic patterns are unpredictable with occasional viral spikes. Which capacity mode balances cost and performance?

  • ○ Provisioned capacity with Auto Scaling
  • ✓ On-demand capacity
  • ○ Provisioned capacity without Auto Scaling
  • ○ Use DynamoDB Accelerator (DAX) only
Answer: B. On-demand capacity
DynamoDB on-demand capacity is ideal for unpredictable workloads with sporadic traffic spikes, as it automatically scales and charges per request without capacity planning. Provisioned capacity with Auto Scaling can handle spikes but requires configuration. Without Auto Scaling, provisioned capacity would be under or over-provisioned. DAX is a cache, not a capacity mode.
Question 10 Domain 1: Design Secure Architectures

A company needs to encrypt data in an Amazon DynamoDB table to meet compliance requirements. How can this be achieved?

  • ✓ Enable encryption at rest using AWS KMS when creating the table.
  • ○ Use client-side encryption before writing data to DynamoDB.
  • ○ Configure SSL/TLS for all queries to the table.
  • ○ Store encryption keys in DynamoDB items.
Answer: A. Enable encryption at rest using AWS KMS when creating the table.
DynamoDB supports encryption at rest using AWS KMS, which can be enabled during table creation. Client-side encryption is possible but adds complexity. SSL/TLS encrypts data in transit. Storing keys in DynamoDB is insecure.
Question 11 Domain 2: Design Resilient Architectures Choose 3

A Solutions Architect is designing a microservices-based application. To ensure the services are loosely coupled and can be scaled independently, which THREE strategies should be adopted? (Select THREE)

  • ○ Use a shared relational database for all services to maintain data consistency.
  • ✓ Implement API Gateway as a single entry point for client requests.
  • ✓ Use asynchronous event-driven communication between services.
  • ○ Deploy all services in a single Amazon ECS cluster with a shared task definition.
  • ✓ Use containerized services that can be deployed independently.
  • ○ Implement a service mesh for service discovery and traffic management.
Answer: B. Implement API Gateway as a single entry point for client requests. | C. Use asynchronous event-driven communication between services. | E. Use containerized services that can be deployed independently.
Options B, C, and E are correct. An API Gateway (B) decouples clients from services. Asynchronous communication (C) prevents cascading failures. Independently deployable containerized services (E) enable independent scaling and deployment. Option A (shared database) creates tight coupling and is an anti-pattern. Option D (single cluster/shared task definition) can create coupling and limit independent scaling. Option F (service mesh) can help with communication but is not a fundamental strategy for loose coupling; it's an implementation tool.
Question 12 Domain 3: Design High-Performing Architectures

A company is deploying a microservices-based application on AWS. Each microservice is deployed as a container. The company wants a fully managed container orchestration service that automatically scales the containers based on demand. Which service should be used?

  • ✓ Amazon ECS with Fargate launch type
  • ○ Amazon EKS with managed node groups
  • ○ AWS App Runner
  • ○ AWS Lambda with container support
Answer: A. Amazon ECS with Fargate launch type
Amazon ECS with Fargate launch type provides a serverless compute engine for containers, automatically managing the underlying infrastructure and scaling containers based on demand. EKS with managed node groups (B) requires managing Kubernetes clusters and node scaling. App Runner (C) is a fully managed service for web applications but is not specifically for microservices orchestration. Lambda with container support (D) runs containers in response to events but is not designed for long-running microservices.
Question 13 Domain 4: Design Cost-Optimized Architectures Choose 2

Which of the following are effective ways to optimize costs for data transfer between AWS services within the same region? (Select TWO.)

  • ○ Use public IP addresses for all inter-service communication.
  • ✓ Use private IP addresses and keep traffic within the same Availability Zone.
  • ✓ Use VPC endpoints for AWS services like S3 and DynamoDB.
  • ○ Use NAT Gateways for all outbound traffic to AWS services.
  • ○ Enable flow logs on all VPCs to monitor traffic.
Answer: B. Use private IP addresses and keep traffic within the same Availability Zone. | C. Use VPC endpoints for AWS services like S3 and DynamoDB.
Data transfer between AWS services in the same Availability Zone (using private IPs) is typically free (e.g., EC2 to EC2, EC2 to RDS). Using VPC endpoints (Gateway or Interface) for services like S3 and DynamoDB allows private connectivity without using the internet, NAT Gateway, or public IPs, avoiding data transfer costs. Using public IPs (A) or NAT Gateways (D) can incur data transfer charges even within a region. Flow logs (E) are for monitoring, not cost optimization.
Question 14 Domain 1: Design Secure Architectures

An application requires secure storage of database credentials that are used by AWS Lambda functions. The credentials must be automatically rotated every 90 days. Which AWS service should be used?

  • ○ AWS Systems Manager Parameter Store
  • ✓ AWS Secrets Manager
  • ○ AWS KMS
  • ○ AWS Certificate Manager
Answer: B. AWS Secrets Manager
AWS Secrets Manager securely stores secrets like database credentials and can automatically rotate them based on a schedule, which meets the requirement. Parameter Store can store secrets but does not natively support automatic rotation. KMS is for encryption key management, and Certificate Manager is for SSL/TLS certificates.
Question 15 Domain 2: Design Resilient Architectures

A three-tier web application is being redesigned for high availability. The current design has web servers in one Availability Zone (AZ) and application servers in another AZ, with a Network Load Balancer (NLB) in front. Users report intermittent failures. How should the architecture be modified to be highly available?

  • ○ Move all web and application servers into the same AZ to reduce latency.
  • ✓ Deploy redundant web and application servers, and the NLB, across at least two AZs.
  • ○ Replace the NLB with an Application Load Balancer (ALB) in a single AZ.
  • ○ Increase the instance size of the web and application servers.
Answer: B. Deploy redundant web and application servers, and the NLB, across at least two AZs.
High availability requires redundancy at every tier across multiple AZs. Deploying all components (web servers, app servers, and the load balancer) across at least two AZs (B) removes single points of failure, aligning with Task 2.2. Consolidating into one AZ (A) increases failure risk. Replacing the NLB with a single-AZ ALB (C) still has a single point of failure. Increasing instance size (D) improves performance but not availability.
Question 16 Domain 3: Design High-Performing Architectures Choose 2

A solutions architect needs to design a highly available and performant network connection between an on-premises data center and multiple Amazon VPCs across different AWS Regions. The connection must provide consistent, high bandwidth and low jitter for real-time data synchronization. Which TWO AWS services should be used? (Select TWO.)

  • ✓ AWS Direct Connect
  • ○ AWS Site-to-Site VPN
  • ○ AWS Global Accelerator
  • ✓ AWS Transit Gateway
  • ○ Amazon Route 53
Answer: A. AWS Direct Connect | D. AWS Transit Gateway
AWS Direct Connect (A) establishes a dedicated, private network connection from on-premises to AWS, providing consistent high bandwidth and low jitter compared to internet-based VPN. AWS Transit Gateway (D) acts as a regional hub to connect multiple VPCs and can be interconnected across regions, and it can be attached to a Direct Connect gateway to provide a single connection point from on-premises to multiple VPCs across regions. Site-to-Site VPN (B) is internet-based and suffers from higher jitter. Global Accelerator (C) improves performance for internet users, not private connections. Route 53 (E) is a DNS service.
Question 17 Domain 4: Design Cost-Optimized Architectures

A company has a three-tier web application hosted on Amazon EC2. The application is deployed in a single region across three Availability Zones. To improve fault tolerance and performance, they want to implement an Active-Active configuration across two AWS regions. What is a cost-effective network design to route user traffic to the nearest healthy region?

  • ○ Use an Amazon CloudFront distribution with origins in both regions.
  • ○ Use Amazon Route 53 with a failover routing policy.
  • ○ Use an Application Load Balancer in each region and AWS Global Accelerator.
  • ✓ Use Amazon Route 53 with a latency-based routing policy and health checks.
Answer: D. Use Amazon Route 53 with a latency-based routing policy and health checks.
Route 53 latency-based routing directs users to the region with the lowest network latency. Combined with health checks, it can route traffic away from an unhealthy region, providing an Active-Active, cost-effective solution. Global Accelerator (C) uses static anycast IPs to direct traffic to the nearest edge location, then over the AWS network to endpoints, but it is not inherently Active-Active across regions without additional configuration. CloudFront (A) is for caching static content, not routing dynamic application traffic. A failover routing policy (B) is for Active-Passive, not Active-Active.
Question 18 Domain 1: Design Secure Architectures Choose 2

A company has compliance requirements to retain data in Amazon S3 for a minimum of 7 years. Which S3 features can help meet this requirement? (Select TWO)

  • ✓ S3 Lifecycle policies
  • ✓ S3 Object Lock
  • ○ S3 Transfer Acceleration
  • ○ S3 Intelligent-Tiering
  • ○ S3 Cross-Region Replication
  • ○ S3 Server-side encryption
Answer: A. S3 Lifecycle policies | B. S3 Object Lock
S3 Lifecycle policies can transition objects to Glacier for long-term retention, and S3 Object Lock prevents deletion or modification for a retention period, ensuring compliance. Transfer Acceleration is for faster uploads, Intelligent-Tiering for cost optimization, Cross-Region Replication for redundancy, and Server-side encryption for security, but not specifically for retention.
Question 19 Domain 2: Design Resilient Architectures Choose 2

Which of the following are benefits of using an Amazon SQS queue between application components? (Select TWO)

  • ○ It eliminates the need for a load balancer.
  • ✓ It decouples components, allowing them to scale independently.
  • ○ It guarantees message delivery in the order they were sent.
  • ✓ It provides a buffer for messages if the consumer is unavailable.
  • ○ It automatically encrypts messages at rest.
Answer: B. It decouples components, allowing them to scale independently. | D. It provides a buffer for messages if the consumer is unavailable.
SQS provides decoupling (B), a core principle for scalable architectures (Task 2.1), and acts as a buffer (D), providing resilience if the consumer fails or is slow. A load balancer (A) is still needed for scaling the consumer services themselves. Only SQS FIFO queues guarantee order (C); standard queues do not. Encryption at rest (E) is an optional feature (SSE) that must be enabled, not automatic.
Question 20 Domain 3: Design High-Performing Architectures

An application requires a persistent, high-performance file system that can be accessed from EC2 instances in multiple AWS Regions. Which service can meet this requirement?

  • ○ Amazon EFS with cross-region replication
  • ○ Amazon FSx for Windows File Server with Multi-Region deployment
  • ○ Amazon S3 with cross-region replication
  • ✓ There is no native AWS service that provides a single shared file system across multiple regions.
Answer: D. There is no native AWS service that provides a single shared file system across multiple regions.
As of the current AWS offerings, there is no native service that provides a single, shared file system (like EFS or FSx) that is natively accessible from EC2 instances in multiple regions with synchronous consistency. EFS is single-region. While S3 can be replicated across regions, it is an object store, not a low-latency file system. Therefore, the most accurate answer is that no native service provides this.

How to use these questions

Every SAA-C03 question contains a constraint, and the constraint is the question. Look for most cost-effective, least operational overhead, highly available or most secure. The correct answer is the one that satisfies the constraint with the least unnecessary complexity.

Two answers are usually technically valid. The differentiator is almost always operational overhead - a managed service beats a self-managed one when the question says least operational overhead, even if the self-managed option is cheaper.

Learn the Well-Architected Framework pillars as a lens rather than a list. Each domain maps to a pillar, and the framing tells you what kind of trade-off the question is testing.

130 minutes for 65 questions is genuinely generous - two minutes each. Use the extra time to re-read the requirement statement before committing, because misreading the constraint is the most common failure mode on this exam.

Frequently asked questions

How hard is the AWS Solutions Architect Associate exam?

It is a real step up from Cloud Practitioner. Questions are long scenario descriptions, two answers are often plausible, and you need 720/1000. Most candidates need 2-3 months of study alongside hands-on AWS work.

How many questions are on SAA-C03?

65 questions in 130 minutes, including multiple-choice and multiple-response items. There is no penalty for guessing.

What is the SAA-C03 passing score?

720 out of 1000 on a scaled score. Because the score is scaled, the raw number of correct answers needed varies between exam versions.

Does SAA-C03 expire?

Yes, after three years. You renew by passing the current exam version or by earning the Professional-level Solutions Architect certification.

Ready to test yourself under exam conditions?

975 questions across 6 timed mock exams, domain-wise drills, and the same scaled scoring as the real thing. Free, no signup.

Related SAA-C03 guides