Amazon Web ServicesLive · 108 questions · 12 notes

SAA-C03: AWS Solutions Architect — Associate

Free practice exam, study notes with architecture diagrams and the official blueprint for AWS Solutions Architect — Associate.

Passing score
720 / 1000
Domains
4
Exam code
SAA-C03

Official SAA-C03 exam blueprint

Domains and weights from Amazon Web Services's exam guide.

DomainWeight
Design Secure Architectures30%
Design Resilient Architectures26%
Design High-Performing Architectures24%
Design Cost-Optimized Architectures20%

Free SAA-C03 sample questions

The first questions of the free diagnostic. Pick your answer, then check it. Why every option is right or wrong, and the What-If twist, open in the workspace.

  1. Question 1 · Design Secure Architectures

    After a penetration test reported that reports-bucket accepted plain HTTP requests, a solutions architect adds this statement to the bucket policy. Other statements already grant the finance role read access, and the change must not break HTTPS clients while it blocks plain HTTP:

    JSON
    {
      "Sid": "DenyInsecureTransport",
      "Effect": "Deny",
      "Principal": "*",
      "Action": "s3:*",
      "Resource": ["arn:aws:s3:::reports-bucket", "arn:aws:s3:::reports-bucket/*"],
      "Condition": {"Bool": {"aws:SecureTransport": "false"}}
    }

    What does the statement enforce for every caller?

    1. AEvery object must be encrypted at rest with SSE-KMS
    2. BEvery request to the bucket must arrive over HTTPS
    3. COnly principals in the bucket owner's account can reach the bucket
    4. DThe bucket can never be made public again
    Check the answer

    Answer: B. A Deny on aws:SecureTransport false is the standard bucket policy pattern that forces TLS for every request.

    Official docs: AWS Documentation: Security best practices for Amazon S3
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
  2. Question 2 · Design High-Performing Architectures

    A web tier behind an Application Load Balancer must scale out before CPU saturation hurts latency and scale in when traffic drops, without anyone maintaining step thresholds or schedules. An engineer applies this policy to the Auto Scaling group:

    Bash
    aws autoscaling put-scaling-policy \
      --auto-scaling-group-name web-asg \
      --policy-name cpu50 \
      --policy-type TargetTrackingScaling \
      --target-tracking-configuration '{
        "PredefinedMetricSpecification": {"PredefinedMetricType": "ASGAverageCPUUtilization"},
        "TargetValue": 50.0}'

    What happens to the group's capacity after this policy is applied?

    1. AIt adds or removes instances to keep the group's average CPU near the target value
    2. BIt changes capacity at fixed times of day, whatever the load
    3. CIt only replaces instances that fail health checks
    4. DIt adds one instance whenever CPU crosses the target and never removes any
    Check the answer

    Answer: A. Target tracking scaling keeps a metric close to a target value by scaling out and in automatically.

    Official docs: AWS Documentation: Target tracking scaling policies for Amazon EC2 Auto Scaling
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
  3. Question 3 · Design Secure Architectures

    In AWS Organizations, administrators in Workloads OU accounts hold AdministratorAccess roles, yet data residency rules require that resources be created only in eu-west-1 and eu-west-3, whatever those roles allow. The platform team drafts this policy statement, excluding global services such as IAM elsewhere in the document:

    JSON
    {
      "Effect": "Deny",
      "NotAction": ["iam:*", "organizations:*", "sts:*", "support:*"],
      "Resource": "*",
      "Condition": {
        "StringNotEquals": {
          "aws:RequestedRegion": ["eu-west-1", "eu-west-3"]
        }
      }
    }

    Where must this statement be attached to bind those administrators?

    1. AIn a service control policy attached to the Workloads OU
    2. BIn an inline policy on each account's AdministratorAccess role
    3. CIn an IAM permissions boundary set by each account's administrators
    4. DIn an S3 bucket policy in the management account
    Check the answer

    Answer: A. Service control policies cap permissions for every principal in member accounts; a Deny on aws:RequestedRegion enforces Region restrictions organization-wide.

    Official docs: AWS Documentation: Service control policies (SCPs)
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
Answer all 10 free SAA-C03 questions

SAA-C03 study notes: key concepts

What each note covers. The full notes, with summaries, examples and memory hooks, open with a free account.

Architecture flow · IAM policy evaluation: explicit deny, SCPs, RCPs, permissions boundaries and session policies

How AWS decides whether a request is allowed
  • API request arrives

      • Explicit Deny in any policy?

        • Yes
          • Request denied

        • No
          • Inside SCP, RCP, boundary and session limits?

            • No
              • Denied by a guardrail or boundary

            • Yes
              • Allowed by identity or resource policy?

                • No
                  • Implicit deny

                • Yes
                  • Request allowed

Drawn fromAWS Documentation: Policy evaluation logic

IAM policy evaluation: explicit deny, SCPs, RCPs, permissions boundaries and session policies

  • An explicit Deny in any applicable policy overrides every Allow, so a NotAction Deny can block far more actions than its author intended.
  • SCPs limit principals in member accounts and RCPs limit resources in member accounts; neither policy type ever grants a permission on its own.
  • A permissions boundary caps a role's identity-based permissions, so an action must be allowed by both the boundary and an identity policy to succeed.
  • Session policies passed on AssumeRole narrow a single session to the intersection of the role's permissions and the session policy document.
  • Cross-account access needs an Allow in the caller's identity policy and in the resource owner's resource-based policy, such as a bucket or key policy.

S3 data protection: encryption choices, Block Public Access, Object Ownership and Object Lock

  • SSE-KMS with a customer managed key gives an editable key policy, per-use CloudTrail logging and cross-account decryption that S3 managed keys never offer.
  • S3 Bucket Keys reduce SSE-KMS cost by using a short-lived bucket-level key, so S3 calls KMS far less often for the same objects.
  • Account-level Block Public Access overrides bucket-level settings, bucket policies and ACLs, so a public policy stays blocked until the account setting changes.
  • Bucket owner enforced Object Ownership disables ACLs and makes the bucket owner the owner of every object, including objects uploaded from other accounts.
  • Object Lock compliance mode prevents any user, including the root user, from deleting or overwriting a locked object version until retention ends.

VPC security: security groups, network ACLs, endpoints, NAT and inspection

  • Security groups are stateful and allow-only, so they cannot deny one address, while network ACLs are stateless and support numbered deny rules.
  • A security group rule can reference another security group as its source, which keeps working when instances are replaced and change addresses.
  • Gateway endpoints for S3 and DynamoDB are free and remove NAT gateway data processing charges, but on-premises clients cannot use them.
  • Interface endpoints place private IP addresses in subnets through PrivateLink, so on-premises networks and other Regions can reach the service privately.
  • A NAT gateway lives in one Availability Zone, so each zone's private subnets should route to their own gateway to survive a zone failure.

High availability and disaster recovery: Multi-AZ, Multi-Region and the four DR strategies

  • Multi-AZ protects against the loss of one Availability Zone inside a Region, while only a Multi-Region design survives the loss of a whole Region.
  • Pilot light replicates data and keeps core infrastructure in the DR Region, but application servers stay off until a disaster is declared.
  • Warm standby runs a smaller but fully working copy of production, so failover only needs scaling and a DNS change within minutes.
  • Multi-site active-active serves users from several Regions at full capacity, giving the fastest recovery at the highest ongoing cost.
  • Route 53 failover routing answers with the primary record while its health check passes and with the secondary record when it fails.

Decoupling: SQS standard and FIFO, dead-letter queues, SNS fan-out, EventBridge and Kinesis

  • An SQS standard queue gives very high throughput with at-least-once delivery, while a FIFO queue adds ordering per message group and deduplication.
  • The visibility timeout must be longer than the processing time, otherwise a message reappears and a second consumer processes it again.
  • A dead-letter queue with a maxReceiveCount moves messages that keep failing aside, so one poison message no longer blocks the rest.
  • SNS fan-out to several SQS queues gives each consumer its own durable copy of every message, with filter policies to select messages.
  • Kinesis Data Streams lets several consumers read the same ordered records independently and replay them within the stream's retention period.

Hybrid and multi-VPC networking: Transit Gateway, peering, Direct Connect, VPN and load balancers

  • VPC peering is point-to-point and non-transitive, so a hub of shared services needs Transit Gateway rather than chains of peering connections.
  • Transit Gateway routes between many VPCs and on-premises connections through route tables, so each new VPC needs only one attachment.
  • A Site-to-Site VPN backing up Direct Connect gives encrypted, low-cost resilience, with BGP keeping the dedicated connection as the preferred path.
  • A Network Load Balancer handles TCP and UDP at very high scale and can use static Elastic IP addresses that partners can allow-list.
  • A Gateway Load Balancer distributes flows to inline security appliances with GENEVE, keeping each flow on the same healthy appliance.
Read the full SAA-C03 study notes

SAA-C03 FAQ

What is the passing score for SAA-C03?

The passing score for AWS Solutions Architect — Associate is 720 / 1000.

Which domains does SAA-C03 cover?

The official exam guide lists 4 domains: Design Secure Architectures (30%); Design Resilient Architectures (26%); Design High-Performing Architectures (24%); Design Cost-Optimized Architectures (20%).

Is there a free SAA-C03 practice test?

Yes. Gyan AI's free diagnostic is 10 SAA-C03 questions with no sign-up, each with an explanation for every option and a link to the official documentation. A free account unlocks the full question set and study notes.

Are these the real SAA-C03 exam questions?

No. Real exam content is confidential under vendor exam agreements, and using leaked "dumps" can get a certification revoked. Gyan AI questions are original and written to the published skills outline.