Governance Day 3
SCS-C03 · Domain 6 · Security Foundations and Governance (14%) · Day 3 of 3

Evaluating compliance: Config, Security Hub CSPM, evidence and reviews

The last day of the last domain. Day 2 made deployments consistent; today you check that everything stays compliant. You will detect noncompliant resources with AWS Config rules, deploy them across accounts, fix them automatically and send alerts, score accounts with Security Hub CSPM, gather evidence with AWS Artifact and Audit Manager, and review designs with the Well-Architected Tool. It ends with a review quiz across all three governance days.

Skill 6.3.1 detect, remediate, notifySkill 6.3.2 collect evidenceSkill 6.3.3 best-practice reviewsDomain 6 review~85 min read + 30 min lab
01

AWS Config rules: detecting noncompliance

Days 1 and 2 were about preventing bad configurations with SCPs, Guard and StackSets. Today is about checking: finding resources that are already wrong, fixing them and proving it to an auditor. Skill 6.3.1 starts with AWS Config. The configuration recorder writes down every change to your resources, and Config rules judge each recorded resource as compliant or not.

Remember the difference from Day 2. CloudFormation Guard and cfn-lint check a template before deployment. A Config rule checks the real resource after deployment, even if someone changed it by hand in the console.

Concept profile

AWS Config rule

often confused with AWS Config recorder / CloudTrail
?What is it?

A check that AWS Config runs against recorded resources. It returns COMPLIANT, NON_COMPLIANT, NOT_APPLICABLE or INSUFFICIENT_DATA for each resource in scope.

⚙How does it work?
  • Three kinds: managed rules written by AWS (owner AWS), custom Lambda rules (owner CUSTOM_LAMBDA) and custom policy rules written in Guard (owner CUSTOM_POLICY), which need no Lambda function.
  • Triggers: configuration change (a resource in scope is created, changed or deleted) or periodic (every 1, 3, 6, 12 or 24 hours; default 24). Some managed rules use both.
  • Guard custom policy rules run on configuration changes; for a scheduled custom check, use a Lambda rule.
  • Evaluation modes: detective (deployed resources) and proactive (planned properties checked with StartResourceEvaluation before deployment; only some rules support it).
  • The configuration recorder must be on and record the resource types the rule needs.
{ }Structure
Change trigger recorder sees a resource created / changed / deleted Periodic trigger every 1, 3, 6, 12 or 24 h (default 24 h) AWS Config rule Managed rule (Owner: AWS) ready-made, set parameters only Custom Lambda rule your code (Python, Java…) Custom policy rule (Guard) no Lambda; change trigger only Evaluation result COMPLIANT NON_COMPLIANT NOT_APPLICABLE INSUFFICIENT_DATA → remediation, events Detective mode checks resources that already exist. Proactive mode checks planned properties before deployment (StartResourceEvaluation). Proactive rules flag; they don’t block.
Two kinds of triggers, three kinds of rules, four possible results.
⇄Config rule vs AWS Config recorder / CloudTrail
Config ruleConfig recorder / CloudTrail
Answers“Is this resource set up correctly?”Recorder: “what does it look like now and before?” CloudTrail: “who called which API?”
OutputCompliance result per resourceConfiguration items / API events
Can trigger a fixYes, through remediationNo (but EventBridge can react to events)
Needs the otherNeeds the recorderRecorder works without rules
Exam hint“detect”, “noncompliant”, “continuously evaluate”“history of changes”, “who made the change”
!Common mistakes the exam tests
  • Expecting a Config rule to block a change: rules detect. Blocking is an SCP, RCP, declarative policy or a proactive Control Tower control.
  • Expecting a proactive rule to stop deployment: proactive evaluation only returns a result; you must act on it (for example in a CloudFormation hook or pipeline).
  • Choosing a change-triggered rule for something that changes without a resource change, such as access key age: that needs a periodic rule (for example access-keys-rotated).
  • Writing a Lambda function when a managed rule already exists (there are hundreds) or when a short Guard rule is enough.
  • Forgetting the recorder: with no recorder or the wrong resource types, rules have nothing to evaluate.
✓Benefits and use cases
  • Continuous checks such as “all S3 buckets block public access” or “no security group allows 0.0.0.0/0 on port 22”.
  • Proof of compliance over time with compliance history.
  • The engine behind Security Hub CSPM controls and Control Tower detective controls.
Think of it like · a school inspector

A school keeps a diary of everything that changes, and an inspector checks the rules.

AWS conceptFamiliar thing
Configuration recorder≈The school diary that notes every change to every classroom
Managed rule≈A standard checklist from the education ministry
Custom Lambda rule≈Your own inspector who can do any check you can describe
Guard custom policy rule≈Your own short checklist written in simple rule sentences
Change trigger≈Inspecting a classroom every time something in it is moved
Periodic trigger≈A visit every 24 hours, even if nothing moved
Real AWS example · a Guard custom policy rule

No Lambda needed. The rule passes only when an S3 bucket has versioning enabled.

$ aws configservice put-config-rule --config-rule '{
  "ConfigRuleName": "s3-versioning-guard",
  "Scope": { "ComplianceResourceTypes": ["AWS::S3::Bucket"] },
  "Source": {
    "Owner": "CUSTOM_POLICY",
    "SourceDetails": [{ "EventSource": "aws.config",
      "MessageType": "ConfigurationItemChangeNotification" }],
    "CustomPolicyDetails": {
      "PolicyRuntime": "guard-2.x.x",
      "PolicyText": "rule check_versioning when resourceType == \"AWS::S3::Bucket\" { supplementaryConfiguration.BucketVersioningConfiguration.status == \"Enabled\" }"
    } } }'

$ aws configservice get-compliance-details-by-config-rule \
    --config-rule-name s3-versioning-guard --compliance-types NON_COMPLIANT \
    --query "EvaluationResults[].EvaluationResultIdentifier.EvaluationResultQualifier.ResourceId"
[ "legacy-logs-bucket" ]

Owner CUSTOM_POLICY tells you it is a Guard rule. The message type shows it runs on configuration changes.

Exam-style scenario

A security engineer must check every 12 hours that no IAM user has an access key older than 90 days, across one account. The team wants the least custom code. Which solution meets the requirement?

  1. A. A custom policy (Guard) rule triggered by configuration changes.
  2. B. The managed rule access-keys-rotated with maxAccessKeyAge 90 and a periodic frequency of 12 hours.
  3. C. A CloudTrail trail with a metric filter on CreateAccessKey.
  4. D. An SCP that denies iam:CreateAccessKey.
Show the worked answer

Answer: B

Key age grows without any configuration change, so the check must be periodic. A managed rule already exists, so no code is needed.

Why not the others

  • A: A change trigger never fires as a key simply gets older.
  • C: Sees new keys, not old ones.
  • D: Prevents new keys, doesn’t find old ones, and may break workloads.
How the exam tests this
  • “Continuously detect misconfigured resources”→ AWS Config rules.
  • “Custom check without managing Lambda code”→ Custom policy rule (Guard).
  • “Check something that changes with time only”→ Periodic trigger (1/3/6/12/24 h).
  • “Check properties before a resource is created”→ Proactive evaluation.
  • “Who changed the security group?”→ CloudTrail (not a Config rule).
Must remember
  • Rules detect; they don’t block.
  • Owners: AWS · CUSTOM_LAMBDA · CUSTOM_POLICY.
  • Periodic: 1, 3, 6, 12, 24 h; default 24.
  • Guard rules: change-triggered, no Lambda.
  • Proactive mode flags before deploy, never blocks.
02

Many accounts: aggregators, organization rules and conformance packs

A real company has dozens of accounts and several Regions. Skill 6.3.1 says “aggregate alerts”. That means two separate jobs. First, push the same rules into every account with organization Config rules or organization conformance packs. Second, pull all results back into one place with an aggregator, so the security team sees one dashboard.

Concept profile

Conformance pack

often confused with Config aggregator
?What is it?

A collection of Config rules and remediation actions, written as one YAML template, that you deploy as a single unit in one account and Region, or across an organization.

⚙How does it work?
  • Start from a sample template (for example operational best practices for HIPAA, NIST 800-53, PCI DSS, CIS) or write your own.
  • Deploy with the console, CLI or an SSM document that stores the template.
  • Organization packs: PutOrganizationConformancePack from the management account or a Config delegated administrator. New accounts that join get the pack automatically; member accounts can’t change or delete it.
  • Organization Config rules work the same way with PutOrganizationConfigRule. Both are API/CLI operations and are Region-specific: repeat per Region.
  • The recorder must already be on in every target account and Region (usually set up by Control Tower).
{ }Structure
Management account or delegated admin PutOrganizationConfigRule PutOrganizationConformancePack API/CLI, one Region per call members can’t edit or delete push (deploy) Account A (prod)eu-west-1, us-east-1Account B (dev)eu-west-1New account Cgets rules automatically pull (replicate) Aggregator many accounts + Regions or a whole organization read-only view can’t deploy or remediate no extra cost Conformance pack = YAML template of Config rules + remediation actions, deployed as one unit. Org accounts need no authorization for an aggregator; individual accounts must authorize it.
Push rules out with organization rules or packs; pull results back with an aggregator.
⇄Conformance pack vs Config aggregator
Conformance packConfig aggregator
DirectionPushes rules into accountsPulls results from accounts
Changes accounts?Yes, creates rules and remediationNo, read-only view
ScopeOne account, or org-wide per RegionMany accounts and Regions in one place
AuthorizationOrg permissions (management or delegated admin)None for org accounts; individual accounts must authorize
Exam hint“deploy the same rules to all accounts”“single view of compliance across accounts and Regions”
!Common mistakes the exam tests
  • Trying to deploy or remediate through an aggregator: aggregators only replicate data and give a read-only view.
  • Running PutOrganizationConfigRule once and expecting all Regions: it is per Region.
  • Forgetting the recorder in member accounts: organization deployment fails or retries for a limited time after a new account joins.
  • Forgetting the service-linked role when deploying from a delegated administrator.
  • Giving every account team rights to edit central rules: organization rules can’t be modified by member accounts, which is the point.
✓Benefits and use cases
  • One compliance dashboard for the whole organization.
  • A HIPAA or PCI DSS starting set of rules in every production account.
  • Guaranteed rules in new accounts without manual work.

Control Tower, Config and the aggregator

If you use AWS Control Tower (Day 1), it turns on the Config recorder in governed accounts, sends Config data to the Log Archive account and creates an organization-wide aggregator in the Audit account (landing zone 3.x). Since landing zone 4.0 (November 2025) the Config integration is optional, Config data goes to its own bucket in the Config aggregator account (the old Audit account), a service-linked Config aggregator replaces the organization aggregator, and member accounts are recorded only in OUs where the AWS Config baseline is enabled. Control Tower detective controls are Config rules under the hood. So on the exam, “Control Tower is in place, now we need extra custom checks in every account” often leads to an organization conformance pack or organization Config rule deployed by the delegated administrator.

Think of it like · a chain of supermarkets

Head office sets rules for every store and wants one report.

AWS conceptFamiliar thing
Conformance pack≈A printed rule book plus a fix-it guide, sent to stores as one package
Organization pack≈Head office mails the rule book to every store, including stores that open next month
Members can’t edit≈Store managers can’t tear pages out of the head office rule book
Aggregator≈The head office dashboard that copies every store’s inspection results
Read-only≈The dashboard can show a dirty fridge but can’t clean it
Per Region≈Each country’s branch must receive its own copy
Real AWS example · deploy a pack to the organization and build an aggregator

Run from the Config delegated administrator account.

$ aws configservice put-organization-conformance-pack \
    --organization-conformance-pack-name baseline-pci \
    --template-s3-uri s3://org-config-packs/Operational-Best-Practices-for-PCI-DSS.yaml \
    --excluded-accounts 444455556666 --region eu-west-1

$ aws configservice put-configuration-aggregator \
    --configuration-aggregator-name org-all \
    --organization-aggregation-source \
      "RoleArn=arn:aws:iam::111122223333:role/ConfigAggregatorRole,AllAwsRegions=true"

$ aws configservice get-aggregate-compliance-details-by-config-rule \
    --configuration-aggregator-name org-all --config-rule-name restricted-ssh \
    --account-id 222233334444 --aws-region eu-west-1 --compliance-type NON_COMPLIANT

The pack is deployed in one Region; repeat the first command for each Region. The aggregator covers all Regions in one call.

Exam-style scenario

A company has 60 accounts in AWS Organizations across three Regions. The security team must (1) make sure the same 25 Config rules exist in every account, including future accounts, without letting account owners delete them, and (2) see compliance for all accounts in one place. Which combination meets these requirements?

  1. A. Create the rules in each account with StackSets and give each team a dashboard.
  2. B. From the Config delegated administrator, deploy an organization conformance pack in each Region, and create an organization aggregator.
  3. C. Create an aggregator and deploy the rules through it.
  4. D. Enable Security Hub CSPM in the management account only.
Show the worked answer

Answer: B

The organization pack reaches new accounts and is protected from member changes; the aggregator gives one read-only view.

Why not the others

  • A: Works but member accounts could change the rules, and there is no single view.
  • C: Aggregators can’t deploy rules.
  • D: One account doesn’t cover the organization, and these are custom rules.
How the exam tests this
  • “Single view of compliance for all accounts and Regions”→ Config aggregator.
  • “Same rules in all accounts, members can’t remove”→ Organization conformance pack / org rules.
  • “Package rules + remediation as one unit”→ Conformance pack (YAML).
  • “Aggregator for non-org accounts”→ Source accounts must authorize it.
  • “Org deployment in two Regions”→ Call the API in each Region.
Must remember
  • Packs/org rules push; aggregators pull.
  • Aggregators are read-only and free.
  • Org rules: management or delegated admin, API/CLI only.
  • Org deployment is per Region.
  • Control Tower detective controls = Config rules.
03

Remediation and notifications

Detection is only half of skill 6.3.1: you must also remediate and notify. AWS Config remediates with AWS Systems Manager Automation runbooks. Notifications travel through Amazon EventBridge, which can send the event to an SNS topic, a Lambda function or another runbook.

Concept profile

AWS Config remediation

often confused with EventBridge + Lambda fix
?What is it?

An SSM Automation document attached to a Config rule. It runs against a noncompliant resource, either when someone chooses Remediate (manual) or automatically.

⚙How does it work?
  • Pick an AWS-owned runbook (for example AWS-DisableIncomingSSHOnPort22, AWSConfigRemediation-ConfigureS3BucketPublicAccessBlock) or write your own.
  • Map parameters: the resource ID parameter is replaced with the noncompliant resource ID at runtime; other parameters are static.
  • Give the runbook an AutomationAssumeRole with only the permissions the fix needs.
  • Automatic remediation runs for every noncompliant resource as soon as it is enabled. Set retries and a time window (for example 5 attempts in 300 seconds); each retry is a paid Automation run.
  • Check failures with describe-remediation-execution-status. Remediation can’t be managed on service-linked rules (for example those created by Security Hub CSPM).
{ }Structure
Config rule resource becomes NON_COMPLIANT automatic SSM Automation runbook AWS-owned or custom document AutomationAssumeRole + RESOURCE_ID Resource fixed re-evaluated; retries if it failed Amazon EventBridge rule Config Rules Compliance Change Security Hub Findings - Imported Targets SNS topic (email, chat) · Lambda (custom fix) · SSM Automation Step Functions (approval workflow) · ticketing via SNS or Lambda Manual remediation: an operator presses Remediate in the console. Automatic: Config runs it for every noncompliant resource as soon as you turn it on. Each retry is a paid Automation run.
The fix path (top) and the notify path (bottom) usually run side by side.
⇄Config remediation vs EventBridge + Lambda fix
Config remediationEventBridge + Lambda fix
TriggerConfig compliance resultAny event pattern (Config, Security Hub, CloudTrail, GuardDuty)
Fix logicSSM Automation runbookYour Lambda code
Retries built inYes (attempts + time window)You build them
Visible in Config consoleYes, per resourceNo
Exam hint“least effort”, “AWS-provided runbook”“custom logic”, “ticket”, “notify and approve”
!Common mistakes the exam tests
  • Enabling automatic remediation on a broad rule without testing: it immediately changes every noncompliant resource, and stale results can even act on resources that are already fixed.
  • Giving the Automation role administrator rights: scope it to the one fix.
  • Expecting the Config rule itself to send email: create an EventBridge rule on Config Rules Compliance Change with an SNS target (or use the Config delivery channel SNS topic for all changes).
  • Trying to attach remediation to a Security Hub CSPM service-linked rule: use Security Hub automation rules or EventBridge instead.
  • Forgetting the SNS topic policy that lets events.amazonaws.com publish.
✓Benefits and use cases
  • Close open SSH ports automatically.
  • Turn on S3 Block Public Access or default encryption the moment it is missing.
  • Alert the owning team in chat or email about every new noncompliant resource.
Think of it like · a building’s fire safety system

Detect, fix and tell people.

AWS conceptFamiliar thing
Config rule≈The smoke detector
Automatic remediation≈The sprinkler that turns on by itself
Manual remediation≈A fire extinguisher that a person must pick up and use
SSM Automation runbook≈The written fire procedure the system follows step by step
Retries≈The sprinkler tries again if the fire is still burning
EventBridge → SNS≈The alarm bell and the text message to the building manager
Real AWS example · auto remediation plus an email alert

Close port 22 automatically, and send an email for every new NON_COMPLIANT result.

$ aws configservice put-remediation-configurations --remediation-configurations '[{
  "ConfigRuleName": "restricted-ssh",
  "TargetType": "SSM_DOCUMENT",
  "TargetId": "AWS-DisableIncomingSSHOnPort22",
  "Parameters": {
    "AutomationAssumeRole": { "StaticValue": { "Values": ["arn:aws:iam::111122223333:role/ConfigFixSSH"] } },
    "SecurityGroupIds": { "ResourceValue": { "Value": "RESOURCE_ID" } } },
  "Automatic": true, "MaximumAutomaticAttempts": 5, "RetryAttemptSeconds": 60 }]'

EventBridge rule pattern (target: SNS topic security-alerts):
{ "source": ["aws.config"],
  "detail-type": ["Config Rules Compliance Change"],
  "detail": { "configRuleName": ["restricted-ssh"],
    "newEvaluationResult": { "complianceType": ["NON_COMPLIANT"] } } }

RESOURCE_ID is replaced with the security group ID of each noncompliant resource.

Exam-style scenario

A company wants S3 buckets that lose Block Public Access to be fixed within minutes, and the bucket owner team must receive an email each time. The solution must use AWS-provided components where possible. What should the security engineer do?

  1. A. Write a Lambda function that scans all buckets every hour and emails the team.
  2. B. Use the managed rule s3-bucket-level-public-access-prohibited with automatic remediation using AWSConfigRemediation-ConfigureS3BucketPublicAccessBlock, and an EventBridge rule for Config Rules Compliance Change that targets an SNS topic.
  3. C. Create an aggregator and subscribe the team to it.
  4. D. Turn on Amazon Macie.
Show the worked answer

Answer: B

Managed rule + AWS runbook + EventBridge/SNS is the standard detect, fix and notify pattern with no custom code.

Why not the others

  • A: Custom code, and hourly is slower than change-triggered.
  • C: Aggregators don’t send notifications or fix anything.
  • D: Macie finds sensitive data; it doesn’t fix bucket settings.
How the exam tests this
  • “Fix noncompliant resources automatically”→ Config automatic remediation with SSM Automation.
  • “Notify a team on noncompliance”→ EventBridge rule → SNS.
  • “Remediation keeps failing”→ describe-remediation-execution-status; check the assume role.
  • “Pass the bad resource to the runbook”→ Resource ID parameter (RESOURCE_ID).
  • “Human approval before the fix”→ Manual remediation, or EventBridge → Step Functions / SSM approval step.
Must remember
  • Remediation = SSM Automation documents.
  • Auto remediation acts on all noncompliant resources at once.
  • Set retries + window; retries cost money.
  • Notify via EventBridge → SNS.
  • No remediation on service-linked rules.

AWS User Notifications: one place for alerts

AWS User Notifications is in the SCS-C03 service list. It is a console service that turns AWS events into notifications without you building EventBridge → SNS → Lambda plumbing in every account. AWS managed notifications (for example AWS Health) arrive by default; notification configurations you create use EventBridge event rules (for example CloudWatch alarm state changes or Health scheduled changes). Delivery channels: the console Notifications Center, email, Amazon Q Developer in chat applications (Slack, Teams) and the AWS Console Mobile App.

With trusted access in AWS Organizations, the management account or a delegated administrator creates one configuration for the whole organization or chosen OUs. Matching events from member accounts create notifications in the admin account only, aggregation can group events (5 minutes or 12 hours), and new accounts are covered automatically.

How the exam tests this
  • “Alert the team by email and chat about events from all accounts, least effort, no custom code”→ AWS User Notifications org configuration from a delegated admin.
  • “Fix the resource automatically”→ Still Config remediation / EventBridge + SSM; User Notifications only notifies.
Must remember
  • User Notifications = notify, not remediate.
  • Org-wide from the management account or delegated admin, needs trusted access.
  • Channels: console, email, chat, mobile app.
04

Security Hub CSPM: standards, controls and central configuration

The blueprint names Security Hub for skill 6.3.1. Be careful with the name. Since 2025 the posture checking service that runs standards and controls is called AWS Security Hub CSPM (Cloud Security Posture Management). A new AWS Security Hub, generally available since December 2025, sits on top and correlates signals from GuardDuty, Inspector and Security Hub CSPM into prioritized risks. Older study material just says “Security Hub” for what is now CSPM.

Concept profile

Security Hub CSPM standards and controls

often confused with AWS Config conformance packs
?What is it?

A service that checks your accounts against security standards (sets of controls), gives a security score, and collects findings from other AWS services and partners in one format (ASFF).

⚙How does it work?
  • Standards include AWS Foundational Security Best Practices (FSBP), CIS AWS Foundations Benchmark, NIST SP 800-53 Rev. 5, PCI DSS and the AWS Resource Tagging Standard.
  • Most controls run as service-linked AWS Config rules, so the Config recorder must be on in every account and Region.
  • Each control (for example S3.1, EC2.19) produces a finding that is PASSED or FAILED, with remediation guidance.
  • Central configuration: integrate with Organizations, pick a delegated administrator and a home Region, then use configuration policies to turn standards and controls on for accounts, OUs or the root; findings from linked Regions flow to the home Region.
  • Automation rules update or suppress findings (severity, workflow status) automatically; findings also go to EventBridge as Security Hub Findings - Imported.
{ }Structure
InputsConfig service-linked rulesGuardDuty · InspectorMacie · Firewall ManagerIAM Access Analyzerpartner productsSecurity Hub CSPMStandards: AWS FSBP, CIS,NIST SP 800-53, PCI DSS,AWS Resource TaggingControls (e.g. S3.1) → findingsSecurity score per standardFindings in ASFF formatAct on itAutomation rules(suppress, set severity)EventBridge → SNS, SSMCross-Region aggregationCentral configurationNaming (2025): the posture service is now AWS Security Hub CSPM.The new AWS Security Hub (GA Dec 2025) correlates these signals into risks.
Findings in, standards and scores in the middle, automation out.
⇄Security Hub CSPM vs AWS Config conformance packs
Security Hub CSPMConfig conformance packs
Built onService-linked Config rulesConfig rules you own
ContentAWS-curated standards and controlsAny rules you choose (sample templates)
RemediationGuidance + EventBridge / automation rulesSSM Automation attached in the pack
Also aggregatesGuardDuty, Inspector, Macie, partner findingsOnly Config results
Exam hint“security score”, “CIS / FSBP”, “central findings”“custom set of rules + remediation”
!Common mistakes the exam tests
  • Enabling Security Hub CSPM but not AWS Config: most controls stay without data.
  • Enabling it account by account: use central configuration with a delegated administrator.
  • Looking for findings from every Region in each Region: use the home (aggregation) Region with linked Regions.
  • Suppressing noisy findings by hand every day: write an automation rule.
  • Mixing up the new Security Hub (risk correlation and exposure) with Security Hub CSPM (standards and controls).
✓Benefits and use cases
  • A posture score against CIS or FSBP for every account.
  • One place for findings from GuardDuty, Inspector, Macie and partners.
  • Org-wide rollout of the same standards with configuration policies.
Think of it like · a hospital health check

Every patient gets the same standard tests and one report.

AWS conceptFamiliar thing
Standard (CIS, FSBP)≈A standard health check package
Control≈One test in the package, such as blood pressure
Security score≈The percentage of tests that came back normal
Findings from GuardDuty, Inspector≈Results sent in from specialist clinics
Automation rule≈A rule that files known harmless results away automatically
Central configuration≈Head office deciding which tests every branch must run
Real AWS example · central configuration policy and an automation rule

From the delegated administrator, in the home Region.

$ aws securityhub create-configuration-policy --name prod-baseline \
  --configuration-policy '{ "SecurityHub": { "ServiceEnabled": true,
    "EnabledStandardIdentifiers": [
      "arn:aws:securityhub:eu-west-1::standards/aws-foundational-security-best-practices/v/1.0.0",
      "arn:aws:securityhub:eu-west-1::standards/cis-aws-foundations-benchmark/v/3.0.0" ],
    "SecurityControlsConfiguration": { "DisabledSecurityControlIdentifiers": ["CloudTrail.6"] } } }'

$ aws securityhub start-configuration-policy-association \
  --configuration-policy-identifier <policy-id> --target '{"OrganizationalUnitId":"ou-prod-abcd"}'

$ aws securityhub create-automation-rule --rule-name sandbox-low --rule-order 1 \
  --criteria '{"AwsAccountId":[{"Value":"444455556666","Comparison":"EQUALS"}]}' \
  --actions '[{"Type":"FINDING_FIELDS_UPDATE","FindingFieldsUpdate":{"Severity":{"Label":"LOW"}}}]'

Check the exact standard ARNs and versions in your Region; the pattern is what the exam tests.

Exam-style scenario

A company with 120 accounts wants every production account checked against the CIS AWS Foundations Benchmark, with new accounts included automatically, and all findings visible from one Region. What should the security engineer do?

  1. A. Enable Security Hub CSPM manually in each account and Region.
  2. B. Designate a delegated administrator, use central configuration with a home Region and linked Regions, and associate a configuration policy that enables CIS with the production OU. Make sure AWS Config is recording in all accounts.
  3. C. Deploy a Config aggregator only.
  4. D. Use AWS Artifact to download the CIS report.
Show the worked answer

Answer: B

Central configuration applies the standard to an OU (including new accounts) and aggregates findings to the home Region; controls need Config.

Why not the others

  • A: Manual, error-prone, misses new accounts.
  • C: An aggregator shows Config data but doesn’t provide the CIS standard and score.
  • D: Artifact has AWS’s own compliance reports, not checks of your accounts.
How the exam tests this
  • “Security score against CIS / FSBP”→ Security Hub CSPM standards.
  • “Enable standards for all accounts in an OU”→ Central configuration + configuration policy.
  • “Controls show no data”→ Turn on the AWS Config recorder.
  • “Automatically lower severity for sandbox findings”→ Security Hub automation rule.
  • “Send failed controls to a ticket system”→ EventBridge (Findings - Imported) → target.
Must remember
  • Name check: Security Hub CSPM = standards and controls.
  • New Security Hub (Dec 2025) correlates risks.
  • Controls need AWS Config.
  • Central config: delegated admin + home Region + policies.
  • Automation rules update or suppress findings.
05

Evidence for auditors: AWS Audit Manager and AWS Artifact

Skill 6.3.2 is about evidence. An auditor asks two different questions. “Is AWS itself secure?” is answered by AWS Artifact, which gives you AWS’s own audit reports and agreements. “Is your use of AWS compliant?” is answered by evidence from your accounts, which AWS Audit Manager collects and organizes automatically.

Concept profile

AWS Audit Manager

often confused with AWS Artifact
?What is it?

A service that continually collects evidence from your accounts, maps it to the controls of a compliance framework, and builds an assessment report for auditors.

⚙How does it work?
  • Choose a prebuilt framework (for example PCI DSS, HIPAA, SOC 2, GDPR, CIS) or build a custom framework from standard and custom controls.
  • Create an assessment: pick the framework, accounts and services in scope, an S3 bucket for reports, and audit owners.
  • Evidence sources: AWS Config and Security Hub CSPM results (compliance checks), CloudTrail (user activity), AWS API calls (configuration snapshots), and manual or uploaded evidence.
  • Delegation: an audit owner sends a control set to a subject matter expert, who reviews evidence, comments and sets control status.
  • Generate an assessment report into your S3 bucket. Audit Manager collects evidence; it doesn’t decide whether you are compliant.
{ }Structure
AWS Artifact evidence about AWS itself Reports: SOC, ISO, PCI, C5 … Agreements: BAA (HIPAA) and others Account or organization agreement Org agreement covers new members Third-party (Marketplace) reports Free, self-service portal AWS Audit Manager evidence about how YOU use AWS Framework (prebuilt or custom) → assessment (scope) Control sets → controls → evidence folders Sources: Config, Security Hub CSPM, CloudTrail, AWS API calls, manual / uploaded evidence Delegation: send a control set to an expert Assessment report in your S3 bucket Collects evidence; does not judge compliance Recent change: Audit Manager is in maintenance mode; no new accounts after 30 April 2026.
Artifact: AWS’s evidence. Audit Manager: your evidence.
⇄Audit Manager vs AWS Artifact
Audit ManagerAWS Artifact
Evidence aboutYour use of AWSAWS’s own infrastructure and controls
ContentConfig, Security Hub CSPM, CloudTrail, API, manual evidenceSOC, ISO, PCI reports; BAA and other agreements
Work neededSet up frameworks and assessmentsDownload or accept, self-service
Status in 2026Maintenance mode, no new accountsFully available, free
Exam hint“collect evidence continuously”, “delegate review”“AWS SOC 2 report”, “sign a BAA”
!Common mistakes the exam tests
  • Sending an auditor to Artifact for evidence about your encryption settings: Artifact only covers AWS’s side of the shared responsibility model.
  • Thinking Audit Manager makes you compliant or gives a pass/fail verdict: it only gathers and organizes evidence.
  • Signing a BAA in each account one by one when an organization agreement from the management account covers all current and future member accounts.
  • Forgetting the delegation feature when the question says “control owners must review evidence for their area”.
  • Ignoring the 2026 change: for new accounts, AWS points to Config conformance packs, Control Tower and Security Hub CSPM instead of Audit Manager.
✓Benefits and use cases
  • Prepare a PCI DSS or HIPAA audit with evidence already organized by control.
  • Let control owners review evidence for their own area.
  • Give auditors AWS’s SOC and ISO reports and manage a HIPAA BAA (Artifact).

AWS Artifact in detail, and the Audit Manager availability change

AWS Artifact is a free self-service portal. Reports include SOC, ISO and PCI attestations and certifications; you accept the Artifact terms before downloading some of them. Agreements include the Business Associate Addendum (BAA), which HIPAA-covered companies usually need before processing protected health information. You can accept an agreement for one account, or as an organization agreement from the management account so that all existing and future member accounts are covered. You can also terminate an agreement later. Artifact also offers third-party reports from AWS Marketplace sellers.

Important 2026 change: AWS Audit Manager is in maintenance mode. From 30 April 2026 it can’t be set up in new accounts, and it won’t get new frameworks or Regions. Existing customers can keep using it, including new assessments. AWS suggests Config conformance packs, the Config resource compliance dashboard, Control Tower and Security Hub CSPM as alternatives. The exam guide still names Audit Manager, so learn its concepts.

Think of it like · buying a house

Two kinds of paperwork.

AWS conceptFamiliar thing
AWS Artifact report≈The builder’s official safety certificate for the building itself
BAA agreement≈A signed contract that the builder will handle your private papers carefully
Organization agreement≈One contract that automatically covers every flat you will ever buy in that building
Audit Manager framework≈The bank’s checklist of documents it needs from you
Evidence collection≈A folder that fills itself with your bills and receipts
Delegation≈Asking your accountant to check the tax section of the folder
Real AWS example · Audit Manager assessment and Artifact agreements

Create an assessment from a framework, then list Artifact agreements.

$ aws auditmanager create-assessment --name pci-2026 \
    --framework-id <pci-dss-framework-id> \
    --assessment-reports-destination destinationType=S3,destination=s3://audit-reports-111122223333 \
    --scope '{"awsAccounts":[{"id":"222233334444"}],"awsServices":[{"serviceName":"s3"},{"serviceName":"kms"}]}' \
    --roles roleType=PROCESS_OWNER,roleArn=arn:aws:iam::111122223333:role/AuditOwner

$ aws auditmanager create-assessment-report --assessment-id <id> --name q4-pci-report

$ aws artifact list-customer-agreements
{ "customerAgreements": [ { "name": "AWS Business Associate Addendum",
    "organizationArn": "arn:aws:organizations::111122223333:organization/o-exampleorg",
    "state": "ACTIVE" } ] }

An organization agreement shows the organization ARN; every member account is covered.

Exam-style scenario

A healthcare startup will store protected health information in a new AWS Organizations environment with 15 accounts, and more accounts will be added. Its compliance officer also needs AWS’s SOC 2 Type II report for a customer due-diligence review. What should the security engineer do?

  1. A. Open a support case to request a BAA and the SOC report.
  2. B. In AWS Artifact, accept the BAA as an organization agreement from the management account, and download the SOC 2 report from Artifact reports.
  3. C. Create an Audit Manager assessment with the HIPAA framework; it signs the BAA.
  4. D. Enable Security Hub CSPM with the NIST standard.
Show the worked answer

Answer: B

Artifact holds both: organization agreements cover current and future member accounts, and AWS’s SOC reports are self-service downloads.

Why not the others

  • A: Not needed; both are self-service in Artifact.
  • C: Audit Manager collects your evidence; it doesn’t create legal agreements (and it isn’t available to new accounts after April 2026).
  • D: Checks your resources, gives no AWS report or BAA.
How the exam tests this
  • “Download AWS’s SOC / ISO / PCI report”→ AWS Artifact reports.
  • “Accept a BAA for all accounts”→ Artifact organization agreement (management account).
  • “Collect evidence mapped to a framework”→ Audit Manager assessment.
  • “Control owners must review evidence”→ Audit Manager delegation of a control set.
  • “Evidence of who did what”→ CloudTrail user activity evidence.
Must remember
  • Artifact = AWS’s evidence; Audit Manager = yours.
  • BAA via Artifact; org agreement covers future accounts.
  • Audit Manager: framework → assessment → control sets → evidence.
  • Audit Manager collects, doesn’t judge.
  • Audit Manager: maintenance mode, no new accounts after 30 Apr 2026.
06

Best-practice reviews: Well-Architected Tool and Trusted Advisor

Skill 6.3.3 asks you to evaluate an architecture against AWS security best practices. Config and Security Hub CSPM check individual resources. The AWS Well-Architected Tool checks the design and processes of a whole workload through guided questions. Trusted Advisor adds automatic checks of your account, including some security checks.

Concept profile

AWS Well-Architected Tool

often confused with AWS Trusted Advisor
?What is it?

A free console service where you define a workload, answer the questions of one or more lenses, and get a list of high and medium risk issues with an improvement plan.

⚙How does it work?
  • The default lens is the AWS Well-Architected Framework with six pillars, including Security. Add lenses from the Lens Catalog (for example Serverless, SaaS, Financial Services) or your own custom lenses for company rules.
  • Answers mark best practices as done or not; missing ones become HRIs (high risk) or MRIs (medium risk).
  • The improvement plan lists what to fix, ordered by risk, with links to guidance.
  • Save a milestone to record the workload’s state at a point in time, and compare later milestones to show progress.
  • Share workloads and custom lenses with other accounts or the organization; Trusted Advisor and Service Catalog AppRegistry data can help answer questions.
{ }Structure
1. Workloadname, Regions,environment,owner2. LensesWA Framework+ catalog lens+ custom lens3. Reviewanswer pillarquestions withthe team4. Riskshigh risk (HRI)medium risk(MRI)5. Planimprovementplan, orderedby risk6. Milestonesnapshot ofthe state;compare laterrepeat after each improvementSecurity pillar areas: security foundations, IAM, detection, infrastructure protection,data protection, incident response, application security. Trusted Advisor data can help answer.
Review, find risks, fix, save a milestone, repeat.
⇄WA Tool vs AWS Trusted Advisor
WA ToolTrusted Advisor
ChecksDesign and processes, through questionsLive account settings, automatically
Who answersYour teamAWS, from your resources
OutputHRIs/MRIs, improvement plan, milestonesCheck status (red / yellow / green) and recommendations
Security contentSecurity pillar: 7 areasSecurity checks: open ports, root MFA, public snapshots…
Exam hint“review architecture”, “track improvement over time”“quick automated checks”, “exposed access keys”
!Common mistakes the exam tests
  • Expecting the WA Tool to scan resources automatically: it is a questionnaire (Trusted Advisor data only helps).
  • Forgetting milestones when the question says “show progress between reviews”.
  • Using the generic lens for company-specific rules: create a custom lens and share it.
  • Thinking every Trusted Advisor check is free: Basic Support gets only selected security and fault tolerance checks plus service limits; full checks and the API need a Business-level plan or higher.
  • Choosing Trusted Advisor when the question needs a compliance standard such as CIS: that’s Security Hub CSPM.
✓Benefits and use cases
  • A security review before a major launch.
  • A company-wide custom lens for internal security standards.
  • Showing auditors and managers that risks went down between milestones.

The Security pillar and Trusted Advisor security checks

The Security pillar has seven design principles: implement a strong identity foundation, maintain traceability, apply security at all layers, automate security best practices, protect data in transit and at rest, keep people away from data, and prepare for security events. Its best practice areas are security foundations, identity and access management, detection, infrastructure protection, data protection, incident response and application security. Trusted Advisor security checks include security groups with unrestricted specific ports, IAM use, MFA on the root account, S3 bucket permissions, public EBS and RDS snapshots and exposed access keys. With Basic Support you must refresh security checks by hand. Trusted Advisor can also show Security Hub CSPM FSBP control results, and it has an organizational view for all accounts.

Think of it like · a yearly car service

Two ways to check a car.

AWS conceptFamiliar thing
Well-Architected review≈A mechanic interview: how do you drive, park and maintain the car?
Lens≈A special checklist for electric cars or taxis
Custom lens≈Your company’s own rules for its delivery vans
HRI / MRI≈“Fix the brakes now” vs “change the tyres soon”
Milestone≈The stamped service book page for this year
Trusted Advisor≈The warning lights on the dashboard that switch on by themselves
Real AWS example · workload review from the CLI

Create a workload with the Framework lens, then read the risk counts and save a milestone.

$ aws wellarchitected create-workload --workload-name payments-api \
    --description "Card payments API" --environment PRODUCTION \
    --aws-regions eu-west-1 --review-owner sec-team@example.com \
    --lenses wellarchitected serverless

$ aws wellarchitected get-lens-review --workload-id 0a1b2c3d4e --lens-alias wellarchitected \
    --query "LensReview.PillarReviewSummaries[?PillarId=='security'].RiskCounts"
[ { "HIGH": 4, "MEDIUM": 3, "NONE": 4, "UNANSWERED": 0, "NOT_APPLICABLE": 0 } ]

$ aws wellarchitected create-milestone --workload-id 0a1b2c3d4e --milestone-name pre-launch-2026-10

Four high risk issues in the Security pillar: fix them, answer again and compare with the next milestone.

Exam-style scenario

A company runs quarterly security reviews of its main workloads. Leaders want to see whether risks are going down over time, and every review must also include internal rules about data classification that AWS guidance doesn’t cover. What should the security engineer use?

  1. A. Trusted Advisor with weekly refresh.
  2. B. The AWS Well-Architected Tool with the Framework lens plus a shared custom lens for the internal rules, saving a milestone after each quarterly review.
  3. C. Security Hub CSPM with the FSBP standard.
  4. D. AWS Artifact reports.
Show the worked answer

Answer: B

Custom lenses add company rules; milestones record each review so progress can be compared.

Why not the others

  • A: Automated account checks; no custom questions or milestone history.
  • C: Resource-level controls, not workload design reviews with company rules.
  • D: AWS’s own audit reports.
How the exam tests this
  • “Evaluate a workload against AWS best practices”→ Well-Architected Tool.
  • “Add the company’s own review questions”→ Custom lens.
  • “Track improvement between reviews”→ Milestones.
  • “Serverless-specific guidance”→ Serverless lens from the Lens Catalog.
  • “Quick automated check for root MFA / open ports”→ Trusted Advisor security checks.
Must remember
  • WA Tool: workload → lenses → HRIs/MRIs → plan → milestone.
  • It is a questionnaire, and free.
  • Custom lens for internal rules; share it.
  • Security pillar: 7 principles, 7 areas.
  • Trusted Advisor full checks need Business-level support.
07

Cheatsheet extras: compliance frameworks and AWS Config vocabulary

Your cheat sheet lists the compliance programs by name (ISO 27001, FedRAMP, HIPAA, NIST, PCI DSS, SAS 70, FIPS 140-2) and the AWS Config terms (configuration item, snapshot, stream, recorder). Exam questions use these words without explaining them, so you must know what each one means and which AWS feature helps. Some cheat sheet facts are out of date; the notes below give today’s behavior.

Concept profile

Compliance programs

often confused with each other (who is audited, and what for)
?What is it?

Standards, laws and certifications that say which security controls an organization must have. AWS is audited against many of them for its side of the shared responsibility model; you download the proof from AWS Artifact and still have to meet them for your side.

⚙How does it work?
  • Certification or report: an outside auditor checks AWS (ISO certificate, SOC report, PCI DSS Attestation of Compliance).
  • Services in scope: each program lists which AWS services were assessed (“AWS Services in Scope by Compliance Program”). Use only in-scope services for regulated data.
  • Agreements: some laws need a contract, for example the HIPAA BAA, accepted in Artifact.
  • Your part: run your own assessment of your workload (the cheat sheet calls it a gap audit); Config conformance packs, Security Hub CSPM standards and Audit Manager frameworks help.
{ }Structure
ProgramWhat it is aboutExam clue → AWS help
ISO/IEC 27001Requirements for an information security management system (ISMS); 27017 adds cloud controls, 27018 protects personal data in the cloud“AWS ISO certificate” → Artifact
SOC 1 / SOC 2 / SOC 3SOC 1: controls that affect customers’ financial reporting (it replaced SAS 70). SOC 2: security, availability, confidentiality… (detailed, under NDA). SOC 3: public summary of SOC 2“auditor needs AWS’s SOC 2 report” → Artifact
PCI DSSProtecting payment card data: 12 requirements (firewalls, no vendor defaults, protect and encrypt card data, anti-malware, secure systems, need-to-know access, unique IDs, physical access, logging, testing, security policy)AWS is a Level 1 service provider; you still need your own assessment; Security Hub CSPM PCI DSS standard, conformance pack
HIPAAUS law protecting health information (PHI)Accept the BAA in Artifact; use HIPAA-eligible services; encrypt PHI
FedRAMPUS government program for assessing and continuously monitoring cloud services“US agency workload” → GovCloud / FedRAMP-authorized services
NIST CSF / SP 800-53US cybersecurity frameworks of controls and best practicesSecurity Hub CSPM NIST 800-53 standard, conformance packs
FISMAUS law requiring federal agencies to secure their information systemsUsually met through FedRAMP
FIPS 140-2 / 140-3US standard that validates cryptographic modules, Levels 1 to 4 (4 highest)FIPS endpoints; KMS and CloudHSM HSMs validated at Level 3
⇄Which one is it?
AuditsTip to tell it apart
ISO 27001The security management system“certificate”, “ISMS”
SOC 1 / 2 / 3Controls of a service provider1 = finance, 2 = security (detailed), 3 = public
PCI DSSCard data handling“credit card”, “cardholder data”
HIPAAHealth data (a law, not a certificate)“PHI”, “BAA”
FIPS 140Crypto modules, not companies“validated cryptographic module”, “Level 3”
!Common mistakes the exam tests
  • Treating AWS’s PCI DSS or SOC report as proof that your workload is compliant.
  • Looking for AWS compliance reports in Config, Security Hub or the AWS Customer Agreement: they are in Artifact.
  • Storing PHI before accepting the BAA, or in a service that isn’t HIPAA eligible.
  • Calling HIPAA a certification: it is a law; AWS offers a BAA and eligible services.
  • Choosing CloudHSM only “because KMS isn’t FIPS validated”: see the note below.
✓Benefits and use cases
  • Answer an auditor quickly with AWS’s own reports.
  • Pick the right Security Hub CSPM standard or conformance pack for your program.
  • Decide which services may hold regulated data.

Older material says “for FIPS 140-2 compliance choose CloudHSM, because KMS is multi-tenant and doesn’t comply”. Today the HSMs behind AWS KMS are FIPS 140 validated at Level 3, and current CloudHSM types are also Level 3. On the exam, CloudHSM is still the answer when the question says single-tenant HSM, you control the keys and users, or PKCS#11 / JCE / CNG interfaces; it isn’t needed just for “FIPS validated”.

AWS Config vocabulary from the cheat sheet

TermMeaning
Configuration item (CI)A point-in-time record of one resource: attributes, relationships, related CloudTrail event IDs
Configuration historyAll CIs of one resource over time (“what did this security group look like last Tuesday?”)
Configuration snapshotA collection of CIs for all recorded resources at one moment, delivered to S3
Configuration streamNew CIs sent to an SNS topic as changes happen
Configuration recorderThe switch that records CIs, per Region; choose all resource types or a list
Delivery channelWhere Config sends history and snapshots: an S3 bucket and optionally an SNS topic
Config roleThe service-linked role AWSServiceRoleForConfig (or your own role that trusts config.amazonaws.com) with read access to resources, plus bucket and topic permissions

Older material says AWS Config has “about 40 managed rules” and a limit of 50 rules. Today there are hundreds of managed rules and much higher quotas. Still true: Config is Regional; turn it on in every Region you use. CloudTrail records Config API calls such as StopConfigurationRecorder, so you can alert when someone turns Config off.

Think of it like · exams at school

Different certificates prove different things.

AWS conceptFamiliar thing
ISO 27001≈The school is accredited: its management system meets the standard
SOC 2 report≈A detailed inspection report, shown only to parents who sign a confidentiality form
SOC 3≈The one-page summary posted on the school website
BAA≈A signed promise from the school nurse about how medical notes are kept
Configuration item≈One photo of a classroom at one moment
Configuration history≈The photo album of that one classroom over the year
Snapshot≈A photo of every classroom taken at the same moment
Real AWS example · Config role trust policy and delivery channel

If you use your own role instead of the service-linked role, it must trust the Config service.

{
  "Version": "2012-10-17",
  "Statement": [{
    "Effect": "Allow",
    "Principal": { "Service": "config.amazonaws.com" },
    "Action": "sts:AssumeRole",
    "Condition": { "StringEquals": { "AWS:SourceAccount": "111122223333" } }
  }]
}

$ aws configservice put-delivery-channel --delivery-channel \
  name=default,s3BucketName=config-history-111122223333,snsTopicARN=arn:aws:sns:eu-west-1:111122223333:config-stream,configSnapshotDeliveryProperties={deliveryFrequency=TwentyFour_Hours}

The role also needs s3:PutObject on the bucket and sns:Publish on the topic (or the bucket policy must allow Config).

Exam-style scenario

A health-tech startup will store patient records on AWS. Its auditor asks for (1) proof of AWS’s own security controls and (2) evidence that the company itself can legally process PHI on AWS. What should the security engineer do?

  1. A. Enable Security Hub CSPM and send the auditor its findings.
  2. B. Download AWS’s SOC 2 and ISO 27001 reports from AWS Artifact, and accept the AWS Business Associate Addendum (BAA) in Artifact before storing PHI in HIPAA-eligible services.
  3. C. Ask AWS Support to run a HIPAA audit of the company’s account.
  4. D. Show the auditor the AWS Customer Agreement.
Show the worked answer

Answer: B

Artifact holds both AWS’s audit reports and the BAA agreement that HIPAA requires.

Why not the others

  • A: Findings are about your resources, not AWS’s controls, and they don’t create the legal agreement.
  • C: AWS doesn’t audit your side of the shared responsibility model.
  • D: The general customer agreement isn’t the HIPAA BAA.
How the exam tests this
  • “Is this AWS service compliant with standard X? Where is the proof?”→ AWS Artifact reports.
  • “Process PHI on AWS”→ Accept the BAA; use HIPAA-eligible services.
  • “Single-tenant HSM, customer-managed users”→ CloudHSM.
  • “What did this resource look like last week?”→ Config configuration history.
  • “Existing IAM role for AWS Config”→ Trust policy for config.amazonaws.com.
Must remember
  • AWS’s reports: Artifact. Your compliance: your own assessment.
  • SOC 1 = finance (old SAS 70) · SOC 2 = security (NDA) · SOC 3 = public.
  • HIPAA = law + BAA; FedRAMP = US government.
  • FIPS 140 Level 3: KMS HSMs and CloudHSM today.
  • Config: CI · history · snapshot · stream · recorder · delivery channel; per Region.
08

Lecture summary map

Day 3 on one screen: detect, fix, notify, prove, review.

1. Mind map

① Config rulesManaged · Lambda · GuardChange vs periodicDetect, don’t block② Many accountsPacks + org rules pushAggregator pullsPer Region, read-only③ Remediate + notifySSM AutomationRetries cost moneyEventBridge → SNS④ Security Hub CSPMFSBP, CIS, NIST, PCINeeds ConfigCentral configuration⑤ EvidenceArtifact: AWS’s sideAudit Manager: yoursBAA org agreement⑥ Best practicesWA Tool + lensesHRIs and milestonesTrusted AdvisorGovernance 3evaluate compliance
Tap a branch to jump back to that part of the lecture. Try redrawing this map from memory on paper.

2. Match the requirement to the service

“Detect noncompliant resources continuously”→AWS Config rules“Custom check without writing Lambda code”→Custom policy rule (Guard)“One view of compliance for all accounts”→Config aggregator (read-only)“Same rules + fixes in every account”→Organization conformance pack“Fix automatically when noncompliant”→Auto remediation (SSM Automation)“Score against CIS / FSBP / PCI DSS”→Security Hub CSPM standards“Auditor wants AWS’s SOC 2 report or a BAA”→AWS Artifact“Review a workload against best practices”→Well-Architected Tool
The left column is the exam wording; the right column is the answer.

3. “If the question says… pick…”

Continuously detect a misconfigurationAWS Config rule
Custom rule, no Lambda to maintainCustom policy rule in Guard
Check that depends on time (key age)Periodic trigger
Same rules in all accounts, members can’t deleteOrganization conformance pack / org rules
One view of all accounts and RegionsConfig aggregator
Fix it automaticallyAuto remediation with SSM Automation
Email the team on noncomplianceEventBridge rule → SNS
CIS / FSBP score for the organizationSecurity Hub CSPM central configuration
Controls show no resultsTurn on the Config recorder
AWS’s SOC 2 report, or a BAAAWS Artifact
Evidence organized by framework controlAudit Manager (existing customers)
Workload review and progress over timeWA Tool + milestones
Store PHI on AWSAccept the BAA in Artifact
Day 3 in one sentence

Detect with Config, fix and tell with Automation and EventBridge, score with Security Hub CSPM, prove with Artifact and Audit Manager, and review the design with the Well-Architected Tool.

09

Must-remember cheat sheet

1. Config rules detect; SCPs prevent.
2. Owners: AWS · CUSTOM_LAMBDA · CUSTOM_POLICY (Guard).
3. Periodic: 1 / 3 / 6 / 12 / 24 h, default 24.
4. Guard rules: change-triggered, no Lambda.
5. Proactive mode flags planned resources; doesn’t block.
6. Conformance pack = YAML of rules + remediation.
7. Org rules/packs: API/CLI, per Region, members can’t edit.
8. Aggregator: read-only, free, can’t remediate.
9. Remediation = SSM Automation + assume role.
10. Auto remediation hits all noncompliant resources at once.
11. Notify: EventBridge → SNS.
12. Name: Security Hub CSPM = standards + controls.
13. CSPM controls need the Config recorder.
14. Artifact: AWS’s reports + BAA; org agreement covers new accounts.
15. Audit Manager: maintenance mode, no new accounts after 30 Apr 2026.
16. WA Tool: lenses, HRIs, milestones; free questionnaire.
17. AWS’s reports: Artifact; SOC 1 = old SAS 70; HIPAA needs a BAA.
18. Config terms: CI · history · snapshot · stream · delivery channel.
19. User Notifications: org-wide alerts from a delegated admin to email, chat or mobile; it doesn’t remediate.
10

Core quiz

Check the basics first. Pick an answer, then press Check.

11

Hard quiz

Harder than most real exam questions on purpose, with one “choose two” and one matching question. Aim for 9 out of 12.

12

Domain 6 review quiz (Days 1–3)

Questions from every governance day. If you miss one, reopen that day’s lecture and its “Must remember” boxes.

13

Practice test questions (from your mock exams)

Eight questions about AWS Config, Trusted Advisor and audit evidence. These questions come from your own mock exams in the practice tests folder, matched to this lecture. Each one was checked against current AWS behavior, so some wording and answers are updated; the source is shown at the end of each explanation. Questions marked (New) were written in October 2026 for exam-guide skills or AWS changes that the mocks don’t cover.

14

Hands-on lab (30 minutes, low cost)

  1. In a sandbox account, open AWS Config and make sure the configuration recorder is on (record at least security groups and S3 buckets). Config charges per configuration item recorded and per rule evaluation, so keep the scope small.
  2. Add the managed rule restricted-ssh. Create a test security group that allows TCP 22 from 0.0.0.0/0 and wait until the rule shows it as NON_COMPLIANT.
  3. Create an SNS topic with your email, then an EventBridge rule on Config Rules Compliance Change for that rule with the topic as target. Re-evaluate the rule and check your email.
  4. Add remediation to the rule with AWS-DisableIncomingSSHOnPort22 and a small role for AutomationAssumeRole. Run it manually on the test group and watch the rule turn compliant.
  5. Open AWS Artifact, accept the terms and download one report (for example a SOC report), and look at the Agreements page without accepting anything. Then create a workload in the Well-Architected Tool (free), answer five Security pillar questions and save a milestone.
  6. Clean up: delete the remediation configuration, the rule, the EventBridge rule, the SNS topic, the test security group and the Automation role. Stop the recorder if this is a personal account. If you tried Security Hub CSPM, disable it before its free trial ends, because control checks are billed per check.