title: "SCS-C03 Decision Cheat Sheet" summary: "A compact decision framework for AWS security detection, response, infrastructure, identity, data protection, and governance scenarios."

How to use the matrices

Cover the recommended mechanism and work from the requirement column. State the trust boundary, owner, failure mode, and evidence before revealing the answer. Then alter one constraint—account ownership, Region, recovery objective, data classification, or latency—and explain whether the decision changes.

The matrices are compression aids, not substitutes for the official blueprint or operational experience. If a row feels like a service-name lookup, add the missing conditions and rejected alternatives.

Final decision sequence

  1. Identify the asset, principal, data, and business outcome.
  2. State the explicit security, compliance, recovery, performance, cost, and operational constraints.
  3. Trace the request, identity, network, data, and evidence paths.
  4. Apply the narrowest control that meets the whole requirement.
  5. Reject alternatives using a named constraint or failure mode.
  6. Explain how the control is deployed, monitored, tested, recovered, and governed.

For incident questions, preserve safety, evidence, and reversibility. For authorization questions, evaluate every applicable policy layer. For data questions, start with ownership and lifecycle. For governance questions, connect requirements to proof and exception ownership.

Signal to detection path

Start from evidence requirements, not a favorite service.

Signal to detection path: Start from evidence requirements, not a favorite service.
NeedPrimary pathValidation
API activityCloudTrail data/management events → analysis and alertingGenerate known action and trace identity, account, Region
Configuration driftAWS Config or approved inventory → rule/evaluationCreate reversible drift and verify evaluation
Threat findingNative/partner finding → Security Hub/EventBridgeRoute a sample and verify ownership
Network behaviorFlow/DNS/firewall/edge logs → analysisGenerate approved traffic and locate the record
Objectives: 1.1, 1.2, 1.3

Incident action order

Preserve optionality and evidence before destructive change.

Incident action order: Preserve optionality and evidence before destructive change.
ConditionActionEvidence
Active credential misuseLimit session/credential while preserving identity trailCloudTrail, identity configuration, session context
Suspicious computeSnapshot/collect then isolate reversiblyVolume/memory where authorized, metadata, network and API logs
Unknown blast radiusQuery organization-wide evidence before declaring containmentAccounts, Regions, resources, identities, timestamps
Objectives: 2.1, 2.2

Infrastructure control selection

Match the control to the actual trust boundary and traffic path.

Infrastructure control selection: Match the control to the actual trust boundary and traffic path.
BoundaryControl familyCommon trap
Internet edgeCloudFront/WAF/Shield/API controlsTreating a security group as an L7 control
Subnet/VPCRoutes, SGs, NACLs, Network Firewall, inspectionIgnoring return paths and centralized routing
AWS service accessPrivate endpoints and endpoint policiesAssuming private DNS alone authorizes access
Compute runtimeWorkload identity, hardened image, vulnerability/runtime controlsUsing long-lived embedded credentials
Objectives: 3.1, 3.2, 3.3

IAM policy layer

Determine effective permissions across every applicable layer.

IAM policy layer: Determine effective permissions across every applicable layer.
RequirementControlLimit
Delegate role maximumPermissions boundaryDoes not grant permissions or constrain resource-policy grants in every context
Organization maximumSCP/RCP as applicableDoes not grant permissions
Resource sharingResource policy and principal trustMust still account for explicit denies and service behavior
Short-lived reductionSession policy and session controlsApplies to that session
Objectives: 4.1, 4.2, 6.1

Encryption and key decision

Choose ownership and lifecycle before choosing a checkbox.

Encryption and key decision: Choose ownership and lifecycle before choosing a checkbox.
NeedPatternEvidence
Provider-managed defaultAWS owned/managed key where control level is sufficientService configuration and access logs
Customer policy/controlCustomer-managed KMS keyKey policy, grants, usage events
External materialImported key material when requiredMaterial lifecycle and backup responsibility
Regional disaster recoveryMulti-Region key only when application design requires itReplica state, policy, rotation behavior
Objectives: 5.1, 5.2, 5.3

Governance and assurance

Connect requirement, preventive control, detection, evidence, and exception ownership.

Governance and assurance: Connect requirement, preventive control, detection, evidence, and exception ownership.
GoalMechanismProof
Account baselineAccount vending/Control Tower/IaCVersioned configuration and enrollment state
Prevent prohibited actionOrganization policy or policy-as-code gateDenied test and change record
Detect driftConfig/inventory/custom evaluationKnown drift produces expected finding
Audit evidenceArtifact/Audit Manager/log archive plus ownershipTraceable control-to-evidence record
Objectives: 6.1, 6.2, 6.3