SaveMyCert
Log in
5 of 5 free questions left today·for 10 a day
DEA-C01 · Domain 4

Data Security and Governance practice questions

Data Security and Governance is worth 18% of the DEA-C01 exam — the lightest of the 4 domains. Applying authentication and authorization, encrypting and masking data, preparing audit logs, and managing data privacy and governance. Official weighting 18%. 6 fully worked examples are further down this page, answers included.

Exam weight
18%
the lightest of the 4 domains
Questions
120
across 5 topics
Free, no account
5/day
a free account allows 10 a day
Explanations
Every option
right and wrong

Build a practice session

5 free questions left today.

Domains

How many?

5 free questions left today.

Mode

Ready when you are

10 fresh questions drawn across 1 of 4 domains, in Learn mode.

Review what you missed

Every question you answer incorrectly, and every question you flag while practising, is saved here automatically. Finish a session and you can come back to re-drill just those.

6 sample Data Security and Governance questions, fully explained

Questions from the DEA-C01 bank mapped to domain 4, with the answer key and the reasoning behind every option. None of them repeat the examples on the main DEA-C01 practice page.

Question 1Data Security and Governance

Data analysts who sign in to AWS account A need to query a data lake owned by account B with Amazon Athena. Account B's security team must define the analysts' data permissions entirely in account B, revoke them with a single change, and avoid creating any IAM users. Which approach meets these requirements?

Choose one.

  • a
    Add account A as a principal in the data lake bucket policy and let the analysts query with their account A credentials.

    A bucket policy grant to another account only works if account A also grants the analysts the S3 permissions, so part of the access is defined outside account B, and Athena's catalog access is not covered.

  • b
    Create an IAM role in account B whose trust policy allows account A principals, and have the analysts assume it with AWS STS. Correct

    The role's permissions live in account B, assuming it yields temporary credentials, and deleting the trust or the role revokes everyone at once.

  • c
    Create an IAM role in account A whose trust policy allows account B principals, and attach the data lake access permissions to it.

    This reverses the direction: a role in account A cannot be granted permissions on account B resources by itself, and account B would not control it.

  • d
    Create an S3 access point in account B whose policy names the analysts' account A identities as principals.

    Access point policies still need matching permissions on the analysts' side in account A, and an access point does not cover Athena's Data Catalog access, so account B does not fully control access.

The concept

Cross-account access to resources is normally granted by a role in the account that owns the resources. The role's trust policy names who may assume it; its permissions policy defines what they can do. Callers use sts:AssumeRole and receive temporary credentials.

Why that’s the answer

The deciding constraints are that account B must hold the whole permission definition and be able to revoke it in one place. A role in account B does exactly that: its permissions policy is the only grant on the data, and changing its trust policy cuts off every analyst. Resource-policy approaches (bucket policy or access point policy naming account A) split the grant across both accounts because the calling identity also needs an allow in its own account. A role in account A is the wrong direction.

How to reason it out
  1. Find the owner of the resources: account B.
  2. Ask where permissions must live: entirely in account B, revocable in one change.
  3. Choose a role in account B trusted by account A, assumed with STS for temporary credentials.

Exam tip: To give another account access you control in one place, create the role in your account and trust theirs.

Applying Authentication Mechanisms: IAM, Secrets Manager, and Networks — the lesson that teaches this.

Question 2Data Security and Governance

A data engineer with full AWS Glue permissions tries to create a Glue job that uses an existing role named GlueETLRole. The request fails with an AccessDenied error that names the iam:PassRole action. An administrator must fix this while following least privilege. What should the administrator do?

Choose one.

  • a
    Grant the engineer iam:PassRole on the ARN of the Glue job being created, with iam:PassedToService set to glue.amazonaws.com.

    PassRole is checked against the role being handed over, not the resource that receives it, so a grant on the job ARN does not cover GlueETLRole.

  • b
    Add the engineer's IAM principal to the GlueETLRole trust policy alongside the Glue service principal.

    The trust policy controls who may assume the role. Glue already trusts it; the missing permission is on the engineer, to hand the role to Glue.

  • c
    Grant the engineer iam:PassRole on the GlueETLRole ARN with iam:PassedToService set to glue.amazonaws.com. Correct

    This is exactly the missing permission, scoped to one role and to the Glue service only.

  • d
    Add an iam:PassRole statement on the GlueETLRole ARN to the permissions policy attached to GlueETLRole itself.

    PassRole is a permission of the principal that passes the role, here the engineer. Granting it to the role being passed changes nothing for the engineer's request.

The concept

When a user configures a service to run with a role (a Glue job role, a Lambda execution role, a CloudFormation service role), AWS checks that the user has iam:PassRole on that role. PassRole is the guardrail that stops users from giving a service more power than they have.

Why that’s the answer

The denied action is iam:PassRole, so the engineer, who hands the role to Glue, needs it. Its Resource is the role being passed (the GlueETLRole ARN), not the Glue job, and the iam:PassedToService condition limits it to glue.amazonaws.com for least privilege. Putting PassRole in the role's own policy grants it to the wrong principal. AssumeRole and trust-policy changes answer a different question: who can become the role.

How to reason it out
  1. Read the error: the denied action is iam:PassRole, a permission of the caller who configures the service.
  2. Identify the resource PassRole is checked against: the role being passed, not the job or resource receiving it.
  3. Distinguish passing a role to a service from assuming the role yourself.
  4. Scope the grant to the single role ARN and to Glue with iam:PassedToService.

Exam tip: Handing a role to a service needs iam:PassRole on that role, scoped by ARN and by iam:PassedToService.

Applying Authentication Mechanisms: IAM, Secrets Manager, and Networks — the lesson that teaches this.

Question 3Data Security and Governance

A 24/7 streaming pipeline signs in to an Amazon Redshift cluster with a database user whose password is in AWS Secrets Manager and is rotated by a Lambda rotation function using the single-user strategy. During each rotation, workers that cached the secret shortly before it changed fail to sign in until they refresh. The rotation schedule must stay in place. Which change prevents these failures?

Choose one.

  • a
    Keep the single-user strategy but shorten the rotation interval so that each cached copy of the password is held for less time.

    Workers still hold a password that stops working the moment the database user is changed; rotating more often creates more failure windows, not fewer.

  • b
    Switch the secret to the alternating-users strategy, so the previous credentials remain valid after each rotation completes. Correct

    Alternating users keeps two database users and changes the one not in current use, so the credentials a worker cached keep working until the next rotation.

  • c
    Set the rotation window to a low-traffic period so that fewer pipeline workers hold a cached copy of the password at rotation.

    A 24/7 pipeline has no idle window; any worker holding the old password still fails, so this reduces the impact at best.

  • d
    Have workers request the AWSPREVIOUS version of the secret whenever a sign-in attempt fails after rotation has run.

    With single-user rotation, the AWSPREVIOUS version holds the old password of the same user, which the database no longer accepts.

The concept

Secrets Manager supports two rotation strategies for database credentials. Single-user rotation changes the password of one user in place, so clients with the old password fail until they refresh. Alternating-users rotation maintains two users and rotates the inactive one, so the previously current credentials stay valid.

Why that’s the answer

The failures come from the single-user strategy itself: the moment the password changes, every cached copy is invalid. Alternating users fixes the cause by updating the other user, leaving the cached credentials valid until the next cycle. Shortening the interval or picking a quieter window only changes how often or how badly the failure happens, and AWSPREVIOUS under single-user rotation points to the same user's old, now-invalid password.

How to reason it out
  1. Identify the strategy in use: single-user, which invalidates the old password immediately.
  2. Ask which strategy keeps a previously issued credential valid during and after rotation.
  3. Reject options that only reduce the frequency or impact of failures.

Exam tip: Clients that cache credentials and cannot tolerate sign-in failures need alternating-users rotation.

Applying Authentication Mechanisms: IAM, Secrets Manager, and Networks — the lesson that teaches this.

Question 4Data Security and Governance

An AWS Glue job uses a Glue connection to reach an Amazon RDS database in private subnets. The VPC has no internet gateway or NAT gateway. The job fails to start with the error "Could not find S3 endpoint or NAT gateway for subnetId". The fix must not add hourly or per-GB processing charges. What should the data engineer do?

Choose one.

  • a
    Create an interface VPC endpoint for Amazon S3 in the connection's subnet and enable private DNS for it.

    An interface endpoint would give a private path to S3, but it carries hourly and per-GB processing charges, which the requirement rules out.

  • b
    Create a gateway VPC endpoint for Amazon S3 and associate it with the route table of the connection's subnet. Correct

    A gateway endpoint for S3 is added as a route-table target, has no hourly or processing charge, and is what Glue looks for in the connection subnet.

  • c
    Create a NAT gateway in a public subnet and add a 0.0.0.0/0 route to it from the route table of the connection's subnet.

    A NAT gateway would satisfy the check, but it bills per hour and per GB processed, and it needs an internet gateway the VPC does not have.

  • d
    Add a self-referencing inbound rule for TCP ports 0-65535 to the security group used by the Glue connection.

    Glue does require this rule so its workers can talk to each other, but a missing rule produces a different error; it gives no path to S3.

The concept

Glue jobs that use a VPC connection run their workers in that subnet and still need to reach Amazon S3 (scripts, temporary files, data). Glue checks for an S3 gateway endpoint or a NAT gateway on the subnet before starting.

Why that’s the answer

The error says Glue found no route to S3 from the connection's subnet. Of the two kinds of endpoint, only the gateway endpoint for S3 is free: it is a route-table entry, not a set of ENIs. The interface endpoint and NAT gateway would both provide a path but break the no-charge constraint. The self-referencing security-group rule is a real Glue requirement but addresses worker-to-worker traffic, not S3 access.

How to reason it out
  1. Read the error: Glue cannot find an S3 path from the subnet.
  2. List what can provide that path: S3 gateway endpoint, S3 interface endpoint, NAT gateway.
  3. Apply the cost constraint: only the gateway endpoint has no hourly or processing charge.

Exam tip: Glue in a private subnet needs an S3 path; the free one is a gateway endpoint on the subnet's route table.

Applying Authentication Mechanisms: IAM, Secrets Manager, and Networks — the lesson that teaches this.

Question 5Data Security and Governance

An AWS Lambda function attached to private subnets must call AWS Secrets Manager to fetch database credentials using the SDK's default endpoint. The VPC has no internet gateway or NAT, and the security team will not allow any route to the internet. Which configuration lets the function retrieve the secret?

Choose one.

  • a
    Create a gateway VPC endpoint for Secrets Manager and add it to the route tables of the private subnets.

    Gateway endpoints exist only for Amazon S3 and Amazon DynamoDB; Secrets Manager is reached through an interface endpoint.

  • b
    Create an interface VPC endpoint for Secrets Manager with private DNS and allow HTTPS from the function's security group. Correct

    Private DNS makes the default Secrets Manager hostname resolve to the endpoint's private IPs, and the endpoint's security group admits HTTPS from the function.

  • c
    Create an interface VPC endpoint for Secrets Manager without private DNS and allow HTTPS from the function's security group.

    Without private DNS, the default Secrets Manager hostname still resolves to public addresses that the function cannot reach, so calls time out unless the code is changed to use the endpoint's own DNS name.

  • d
    Launch a NAT gateway in a public subnet and route the private subnets to it so the function reaches the public endpoint.

    This adds a route to the internet (through an internet gateway), which the security team does not allow.

The concept

Most AWS services, including Secrets Manager, are reached privately through interface VPC endpoints (AWS PrivateLink): ENIs with private IPs in your subnets, protected by a security group. Private DNS lets unchanged SDK code use them.

Why that’s the answer

Two facts decide this. Secrets Manager has no gateway endpoint, so the private path must be an interface endpoint. And because the function uses the SDK's default endpoint, private DNS must be enabled so that the regular service hostname resolves to the endpoint's private IP addresses; without it the call still goes to the public hostname and times out. The endpoint's security group must also allow inbound HTTPS from the function. A NAT gateway would work technically but breaks the no-internet-route rule.

How to reason it out
  1. Rule out gateway endpoints: they exist only for S3 and DynamoDB.
  2. Choose an interface endpoint for Secrets Manager.
  3. Check DNS: unchanged SDK code needs private DNS on the endpoint.
  4. Allow HTTPS (443) from the function to the endpoint security group.

Exam tip: Interface endpoint plus private DNS lets unchanged SDK calls reach a service privately; gateway endpoints are only for S3 and DynamoDB.

Applying Authentication Mechanisms: IAM, Secrets Manager, and Networks — the lesson that teaches this.

Question 6Data Security and Governance

A company's analytics VPC reaches Amazon S3 through a gateway VPC endpoint. The security team wants workloads in the VPC to be able to reach only the company's approved data lake buckets through that path, so that a workload cannot copy data to a bucket in some other account even if its IAM permissions would allow it. Where should this restriction be configured?

Choose one.

  • a
    In outbound security group rules on the workloads that use the S3 managed prefix list as the destination.

    The S3 prefix list covers the IP ranges of every bucket in the Region, so security groups cannot tell the approved buckets from any other bucket.

  • b
    In network ACL rules on the private subnets that deny traffic to the IP ranges of non-approved buckets.

    S3 buckets share the same service IP ranges, so network ACLs cannot select individual buckets.

  • c
    In the VPC endpoint policy, allowing S3 actions on the ARNs of the approved buckets and their objects. Correct

    The endpoint policy is evaluated for every request through the endpoint and can limit it to specific bucket ARNs, whatever the caller's IAM policy allows.

  • d
    In the bucket policies of the approved buckets, denying requests unless aws:SourceVpce equals the endpoint ID.

    This controls who can reach the approved buckets, but it places no limit on requests from the VPC to other buckets, such as a bucket in another account.

The concept

A VPC endpoint policy is a resource-based policy on the endpoint. It is evaluated together with identity and bucket policies for every request through the endpoint, so it can cap what the network path can be used for.

Why that’s the answer

The requirement is about where traffic from the VPC can go, not about who can reach the approved buckets. Only the endpoint policy is evaluated for every S3 request leaving through the endpoint, and listing the approved bucket ARNs there blocks requests to any other bucket. Bucket policies with aws:SourceVpce are the tempting near-twin: they lock the approved buckets to the VPC, but a bucket in another account is governed by its owner's policy. Security groups and network ACLs work on IP addresses, which S3 buckets share.

How to reason it out
  1. Restate the goal: limit destinations reachable from the VPC (data exfiltration), not callers of the approved buckets.
  2. Rule out IP-based controls: buckets share S3 service IP ranges.
  3. Pick the policy evaluated on every request through the endpoint: the endpoint policy.

Exam tip: Endpoint policy limits where traffic from your VPC can go; aws:SourceVpce in a bucket policy limits where traffic to your bucket can come from.

Applying Authentication Mechanisms: IAM, Secrets Manager, and Networks — the lesson that teaches this.

What DEA-C01 domain 4 tests, topic by topic

The official exam guide breaks Data Security and Governance into 5 topics. The question bank follows the same split, so a weak topic shows up as a cluster of misses you can go back and read.

Published DEA-C01 practice questions per topic in Data Security and Governance
TopicWhat it coversQuestions
Apply authentication mechanismsOfficial DEA-C01 task statement (guide v1.1). Updating VPC security groups; creating and updating IAM groups, roles, endpoints, and services; creating and rotating credentials (AWS Secrets Manager); setting up IAM roles for access (Lambda, Amazon API Gateway, AWS CLI, CloudFormation); applying IAM policies to roles, endpoints, and services (S3 Access Points, AWS PrivateLink); managed vs unmanaged services; using domains, domain units, and projects for SageMaker Unified Studio.24
Apply authorization mechanismsOfficial DEA-C01 task statement. Creating custom IAM policies when managed policies fall short; storing application and database credentials (Secrets Manager, AWS Systems Manager Parameter Store); providing database users, groups, and roles access (Amazon Redshift); managing permissions through AWS Lake Formation (Redshift, EMR, Athena, S3); authorization methods (role-based, tag-based, attribute-based); constructing least-privilege policies.24
Ensure data encryption and maskingOfficial DEA-C01 task statement. Applying data masking and anonymization per compliance laws or company policy; using encryption keys to encrypt/decrypt data (AWS KMS); configuring encryption across AWS account boundaries; enabling encryption in transit or before transit.24
Prepare logs for auditOfficial DEA-C01 task statement. Tracking API calls with AWS CloudTrail; storing application logs with Amazon CloudWatch Logs; centralized logging queries with AWS CloudTrail Lake; analyzing logs (Amazon Athena, CloudWatch Logs Insights, Amazon OpenSearch Service); integrating services for logging (Amazon EMR for large log volumes).24
Understand data privacy and governanceOfficial DEA-C01 task statement (guide v1.1). Granting permissions for data sharing (Amazon Redshift data sharing); implementing PII identification (Amazon Macie with Lake Formation); preventing backups/replications to disallowed AWS Regions; viewing configuration changes (AWS Config); maintaining data sovereignty; managing data access through SageMaker Catalog projects; governance data frameworks and data-sharing patterns.24
Total120

Revise Data Security and Governance before you drill it

Other DEA-C01 domains

Data Security and Governance: your questions

Data Security and Governance is domain 4 of the DEA-C01 exam guide and carries 18% of the scored content — the lightest of the 4 domains. On a 65-question paper that works out to roughly 12 questions, though AWS does not publish an exact per-domain count and individual exam forms vary.

Source

The domain weight and topic list on this page come from the official DEA-C01 exam guide.