SOA-C03 — AWS Certified CloudOps Engineer – Associate Cheat Sheet

Compact AWS SOA-C03 Cheat sheet for service selection, monitoring, automation, security, networking, reliability, and troubleshooting.

Use the tables for a quick pre-exam check. Expand a topic’s notes for explanations, examples, and additional distinctions.

Scope and study context

This page supports IT Mastery exam prep with original practice questions. It is not affiliated with AWS.

Exam-use orientation

This independent Cheat Sheet supports preparation for the AWS Certified CloudOps Engineer – Associate (SOA-C03) exam from AWS. Use it as a scenario decision guide: the exam often tests which AWS service, operational control, or troubleshooting step best fits a production operations problem.

CloudOps thinking pattern

Question asks about…First decide…Then choose based on…
MonitoringMetric, log, trace, event, or audit record?CloudWatch, X-Ray, EventBridge, CloudTrail, AWS Config
AutomationOne-time command, recurring desired state, patching, or workflow?AWS Systems Manager capability or AWS CloudFormation
Change managementInfrastructure template, app deployment, or instance replacement?CloudFormation, CodeDeploy, Auto Scaling instance refresh
ReliabilityHA in one Region or DR across Regions?Multi-AZ, backups, replication, Route 53 failover
SecurityIdentity, encryption, detection, or compliance evidence?IAM, KMS, CloudTrail, Config, GuardDuty, Security Hub
NetworkingRouting, DNS, firewalling, private access, or edge delivery?VPC route tables, Route 53, security groups/NACLs, VPC endpoints, CloudFront
Cost/performanceRightsizing, purchasing, data transfer, or storage tiering?Compute Optimizer, Cost Explorer, Budgets, lifecycle policies
Notes and examples

Exam habit: eliminate answers that are manually operated, not highly available, not least privilege, or do not produce auditable operational evidence.

Core AWS operations service-selection matrix

Operational needPreferWhyCommon trap
Audit who called AWS APIsAWS CloudTrailRecords management events and optional data eventsCloudWatch Logs show app/system logs, not complete API audit history
Detect resource configuration drift/complianceAWS ConfigTracks resource configuration history and evaluates rulesCloudTrail tells who changed something, not whether current state is compliant
Alarm on metric thresholdAmazon CloudWatch alarmNative metric evaluation and actionsEventBridge is for event patterns, not continuous metric evaluation
Route AWS service events to targetsAmazon EventBridgeEvent bus, rules, schedules, SaaS/custom eventsCloudWatch alarm actions are limited to alarm state transitions
Centralize application logsCloudWatch LogsLog groups, retention, metric filters, Logs InsightsCloudTrail is not an application log platform
Run commands on managed instancesSystems Manager Run CommandRemote command execution without inbound SSH/RDPRequires SSM Agent, IAM role, and network path to Systems Manager endpoints
Enforce recurring instance configurationSystems Manager State ManagerMaintains desired state associationsRun Command is better for ad hoc execution
Patch EC2 or hybrid nodesSystems Manager Patch ManagerBaselines, maintenance windows, patch complianceUser data is not patch management
Secure shell access without opening portsSystems Manager Session ManagerAuditable sessions through SSMStill requires IAM permissions and managed instance connectivity
Automate operational runbookSystems Manager AutomationStep-based remediation workflowsLambda is useful for code, but Automation has runbook-native actions
Provision infrastructure as codeAWS CloudFormationDeclarative stacks, change sets, drift detectionCLI-created resources are harder to audit and reproduce
Deploy application revisionsAWS CodeDeployIn-place/blue-green deployment strategiesCloudFormation manages infrastructure; CodeDeploy manages app rollout
Replace Auto Scaling instances safelyEC2 Auto Scaling instance refreshGradual replacement using launch template/config changesUpdating the launch template alone does not replace existing instances
Central backup policyAWS BackupCross-service backup plans and vaultsSnapshots alone do not provide centralized policy/compliance views
Private access to AWS servicesVPC endpointsAvoid public internet paths for supported servicesNAT gateway provides outbound internet, not private service access
Edge caching and TLS termination near usersAmazon CloudFrontGlobal CDN, caching, origin protection optionsRoute 53 does DNS routing; it does not cache content
Detect suspicious account or workload activityAmazon GuardDutyThreat detection from logs and signalsSecurity Hub aggregates findings; it is not the primary detector
Aggregate security postureAWS Security HubConsolidates findings and standards checksConfig rules are resource compliance checks, not a findings hub
Discover sensitive data in S3Amazon MacieS3 data discovery and classificationS3 Inventory lists objects; it does not classify sensitive content
Analyze IAM external accessIAM Access AnalyzerIdentifies resource policies allowing external accessIAM credential report is about users/passwords/keys
Store database credentials with rotationAWS Secrets ManagerManaged secret lifecycle and rotation integrationParameter Store can store secrets, but rotation is not the same feature set
Store config parametersSystems Manager Parameter StoreHierarchical config values, optional encryptionDo not hard-code config in AMIs or user data
Manage encryption keysAWS KMSKey policies, grants, envelope encryption integrationIAM permission alone may not be enough; key policy must allow use
View AWS account health eventsAWS HealthService and account-specific operational eventsCloudWatch health checks monitor endpoints, not AWS account advisories
Govern accounts at scaleAWS OrganizationsConsolidated management, SCP guardrailsSCPs limit permissions; they do not grant permissions

Monitoring, logging, and event response

Observability services: high-yield distinctions

Service/featureBest forExam cuesNot best for
CloudWatch metricsNumeric time-series performance and healthCPU, latency, errors, queue depth, custom app metricFull audit trail or config history
CloudWatch alarmsThreshold, anomaly, or metric math alarm actionsNotify, scale, recover, stop, route incidentComplex event enrichment
CloudWatch LogsCentral log storage and searchApplication logs, OS logs, Lambda logs, VPC flow logs destinationLong-term object archive unless exported/archived
CloudWatch Logs InsightsAd hoc log query and troubleshootingFilter errors, aggregate by field, recent incident analysisPermanent business analytics warehouse
CloudWatch metric filtersTurn log patterns into metricsCount “ERROR” strings, unauthorized attemptsFree-form historical log analytics
CloudWatch AgentOS/process/custom metrics and logs from EC2/on-premMemory, disk, swap, app logsNative AWS service metrics that already exist
EventBridgeMatch events and route to targetsEC2 state change, scheduled automation, SaaS/custom busContinuous metric thresholding
CloudTrailAPI activity auditWho changed security group? Who deleted object?Instance CPU/memory monitoring
AWS ConfigResource inventory, config history, complianceIs S3 public access blocked? Has SG changed?User login/session troubleshooting
X-RayDistributed tracingService map, trace latency, segment errorsInfrastructure patch compliance
Notes and examples

CloudWatch alarm decision points

DecisionChoose this when…Notes
Standard metric alarmOne metric or metric math expression is enoughMost common alarm scenario
Composite alarmNeed to reduce noise by combining alarm statesUse when multiple symptoms must be true before paging
Anomaly detectionNormal baseline varies over timeGood for cyclical traffic patterns
Treat missing data as breachingMissing metric is itself a failureUseful for heartbeat/custom metrics
Treat missing data as not breachingSilence can be normalAvoid false alarms for sparse metrics
Metric mathNeed derived signalExample: error percentage from errors and requests
Detailed monitoring/custom metricsNeed more granular or non-default dataMemory and disk require agent/custom metrics on EC2
Alarm action to Auto ScalingNeed scaling responseScaling policy should align with application behavior
Alarm action to SNS/EventBridge/Incident ManagerNeed notification or workflowUse structured incident routing for operations teams

CloudWatch Logs Insights patterns

fields @timestamp, @message
| filter @message like /ERROR|Exception|Timeout/
| sort @timestamp desc
| limit 50
fields @timestamp, @logStream, status, latency
| filter status >= 500
| stats count(*) as errors, avg(latency) as avgLatency by bin(5m)
| sort bin(5m) desc

Event-driven remediation pattern

Event sourceMatch withTarget examplesUse case
EC2 instance state changeEventBridge ruleLambda, Systems Manager Automation, SNSReact to stopped/terminated instances
AWS Health eventEventBridge ruleSNS, Incident Manager, ticket workflowNotify on account-specific AWS events
CloudTrail API eventEventBridge ruleLambda, Step Functions, SNSDetect high-risk API calls quickly
Scheduled eventEventBridge schedule/ruleSystems Manager Automation, LambdaRun maintenance tasks
Config compliance changeConfig rule/EventBridgeAutomation, SNSRemediate noncompliant resources

Automation, provisioning, and change management

CloudFormation operations reference

NeedCloudFormation featureExam note
Preview stack changesChange setSafer than direct update for production
Detect manual changesDrift detectionIdentifies resources that no longer match template where supported
Protect critical resource from replacement/deletionDeletionPolicy, UpdateReplacePolicy, stack policyUse Retain or Snapshot where appropriate
Reuse common templatesNested stacks/modulesGood for standardized patterns
Deploy to multiple accounts/RegionsStackSetsFits organization-scale rollout
Pass values between stacksOutputs and exportsAvoid hard-coded IDs
Create resources conditionallyConditionsUseful for environment-specific resources
Bootstrap EC2 on createUser data, cfn-init, cfn-signalSignals help CloudFormation wait for successful configuration
Roll back failed updateAutomatic rollback or continue update rollbackKnow how to recover stacks stuck during failed updates
Manage IAM resourcesCapabilities acknowledgmentIAM creation often requires explicit deployment capability
Notes and examples
aws cloudformation validate-template \
  --template-body file://template.yaml

aws cloudformation deploy \
  --template-file template.yaml \
  --stack-name app-prod \
  --capabilities CAPABILITY_NAMED_IAM

aws cloudformation detect-stack-drift \
  --stack-name app-prod

Systems Manager capability map

CapabilityBest forRequires/depends on
Fleet ManagerInventory and manage nodesManaged instances
Session ManagerShell access without inbound portsSSM Agent, IAM, endpoint/internet connectivity
Run CommandExecute commands at scaleManaged instance role and target selection
State ManagerKeep configuration in desired stateAssociations and documents
Patch ManagerPatch baselines and complianceMaintenance windows optional but common
AutomationMulti-step runbooksIAM service role/permissions
DistributorInstall software packagesPackage definitions
Parameter StoreApp configuration and secure stringsKMS for encrypted secure strings
InventoryCollect software/config metadataSSM Agent and association
Maintenance WindowsScheduled operational tasksRegistered targets and tasks
OpsCenterTrack operational issuesIntegrates with alarms/events
Change ManagerControlled change workflowsApproval and change templates
aws ssm send-command \
  --document-name "AWS-RunShellScript" \
  --targets "Key=tag:Role,Values=web" \
  --parameters commands='["uptime","df -h"]'

Deployment choices

ScenarioPreferWhy
Deploy new Lambda version graduallyCodeDeploy with Lambda deployment configSupports traffic shifting and rollback
Deploy app to EC2 fleetCodeDeployLifecycle hooks and deployment groups
Replace EC2 instances using new launch templateAuto Scaling instance refreshOperationally simple fleet replacement
Manage immutable infrastructureCloudFormation + AMI/launch template + Auto ScalingReproducible state
Blue/green container deploymentECS deployment controller/CodeDeploy depending setupSafer traffic shifting
Manual emergency config changeSystems Manager Automation/Run CommandAuditable and repeatable
Infrastructure resource updateCloudFormation change setAvoid console drift
Complex orchestration across servicesStep Functions or Systems Manager AutomationChoose based on app workflow vs ops runbook

Systems Manager operations

Systems Manager is frequently the most operationally correct answer when the scenario asks for controlled fleet management without opening inbound administrative ports.

Systems Manager featureUse it for
Session ManagerBrowser/CLI shell access without SSH/RDP inbound exposure.
Run CommandExecute commands across managed nodes.
Patch ManagerPatch baselines, patch groups, and maintenance windows.
State ManagerMaintain desired configuration over time.
AutomationMulti-step operational runbooks and remediation.
InventoryCollect software and configuration inventory.
Parameter StoreStore configuration and secrets-like values.
OpsCenterTrack and manage operational issues.
DistributorPackage and distribute software agents.
Maintenance WindowsSchedule disruptive or controlled operations.

SSM managed-node checklist

If an instance is not showing as a managed node, check:

  • SSM Agent installed and running.
  • IAM role attached with required Systems Manager permissions.
  • Network path to Systems Manager endpoints through internet, NAT, or VPC interface endpoints.
  • Correct AWS Region.
  • Supported operating system and instance state.
  • No restrictive proxy, DNS, endpoint policy, or security control blocking communication.

Compute, scaling, and load balancing

EC2 operational troubleshooting

SymptomCheck firstLikely direction
Instance unreachableSecurity group, NACL, route table, public/private IP, SSM statusSeparate network path problem from OS problem
System status check failedAWS host/network issueStop/start, recover, or allow AWS remediation depending scenario
Instance status check failedGuest OS/app issueCheck boot logs, CPU, disk, networking config
User data did not workCloud-init logs, script syntax, IAM role, network accessUser data normally runs at first boot unless configured otherwise
Cannot access S3 from private subnetRoute/NAT or S3 VPC endpoint policyPrefer gateway endpoint for private S3 access where appropriate
App lost AWS permissionsInstance profile, role policy, SCP/permission boundary, STS credentialsTemporary credentials come from role metadata
Memory/disk alarm missingCloudWatch Agent/custom metricsDefault EC2 metrics do not include all OS-level metrics
Replacement instance not configuredAMI, launch template, user data, SSM State ManagerAvoid snowflake instances
Notes and examples

Auto Scaling decisions

NeedFeatureNotes
Maintain fixed capacityDesired/min/max capacityHealth checks replace failed instances
Scale around target metricTarget tracking policyCommon for CPU, request count, custom utilization metric
Scale by thresholds/stepsStep scalingUseful when response should vary by severity
Scale on scheduleScheduled scalingGood for predictable business hours
Prepare for future demandPredictive scalingUse when historical patterns are reliable
Let instances finish work before terminationLifecycle hooksPair with Lambda/SNS/SQS/Systems Manager
Use load balancer healthELB health checks in Auto ScalingReplaces instances failing app-level checks
Safely roll new launch templateInstance refreshCombine with health checks and warmup
Keep scale-in from killing special nodeInstance protectionUseful for stateful/critical instances, but avoid permanent snowflakes

Load balancer selection

Load balancerChoose forKey featuresAvoid when…
Application Load BalancerHTTP/HTTPS appsHost/path routing, redirects, header rules, WebSocket, target groupsNeed static IP at L4
Network Load BalancerTCP/UDP/TLS, high performance, static IP needsLow latency, source IP preservation patterns, TLS passthrough/terminationNeed advanced HTTP routing
Gateway Load BalancerThird-party virtual appliancesTransparent inspection with appliancesNormal web app load balancing
Classic Load BalancerLegacy workloadsOlder EC2-era optionNew architectures should usually choose ALB/NLB

ALB/NLB troubleshooting

ProblemCheck
Targets unhealthyTarget security group, health check path/port/protocol, app listener, NACL, target response code
502/503 errorsTarget availability, listener rules, target group health, backend timeouts
Client IP handlingALB uses headers; NLB can preserve source IP in supported patterns
TLS issueCertificate in ACM/IAM, listener protocol, SNI, security policy
Sticky sessions requiredALB target group stickiness or app-level session design
Slow scale-in connection dropsDeregistration delay and app graceful shutdown

EC2 operational review

TopicHigh-yield point
AMIsUse for repeatable instance builds and faster recovery.
User dataGood for bootstrapping; avoid making it the only place critical configuration exists.
Instance profilesRequired for EC2 to assume an IAM role.
System status check failureOften AWS infrastructure-related; stop/start or recovery may help depending on root volume and instance type.
Instance status check failureOften OS, network config, exhausted resources, or boot issue.
EBS-backed instanceCan generally be stopped and started.
Instance storeEphemeral; data is lost when the instance stops, terminates, or underlying disk fails.
EBS snapshotsIncremental backups stored in S3-managed infrastructure.
Elastic IPStatic public IPv4 address; watch for unnecessary allocation and cost.

Auto Scaling groups

FeatureUse it forCandidate trap
Launch templateDefines instance configurationPrefer launch templates over older launch configurations.
Desired capacityCurrent intended number of instancesScaling policies adjust desired capacity.
Minimum / maximumGuardrails for scaling rangeToo-low max can block scale-out.
Target trackingKeep a metric near a targetUsually preferred for common scaling needs.
Step scalingDifferent adjustments by breach sizeMore control, more tuning.
Scheduled scalingKnown time-based demandNot reactive to unexpected spikes.
Predictive scalingForecast-based scalingNeeds predictable patterns.
Lifecycle hooksRun actions before launch/termination completesUseful for registration, draining, or cleanup.
Health checksReplace unhealthy instancesELB health checks can be used in addition to EC2 checks.
Warm poolsReduce scale-out timeAdds operational and cost considerations.

Load balancer selection

NeedChoose
HTTP/HTTPS routing, host/path rules, redirectsApplication Load Balancer
TCP/UDP/TLS, very high performance, static IP supportNetwork Load Balancer
Third-party virtual appliance insertionGateway Load Balancer
HTTP target health and path-based routingALB target groups
Preserve client source IP at L4NLB patterns
Blue/green or canary with target group shiftingALB / CodeDeploy patterns

Deployment strategies

StrategyUse whenTradeoff
In-placeUpdate existing resourcesLower resource cost, higher rollback risk.
RollingReplace graduallySome mixed-version period.
Blue/greenShift traffic to a separate new environmentMore resources, cleaner rollback.
CanarySend small traffic percentage firstGood risk control; requires monitoring.
ImmutableReplace infrastructure rather than mutate itStrong consistency; requires automation.

Storage and database operations

Amazon S3 operations

NeedFeatureExam note
Block public exposureS3 Block Public Access + bucket policy reviewAccount-level and bucket-level controls matter
Audit object-level API accessCloudTrail data eventsManagement events alone do not show every object operation
Monitor bucket complianceAWS Config rulesGood for encryption, public access, versioning checks
Recover deleted/overwritten objectsVersioningLifecycle can manage old versions
Replicate objectsSame-Region or Cross-Region ReplicationVersioning is required for replication
Enforce encryptionDefault encryption and bucket policyKMS permissions must allow use when SSE-KMS is selected
Archive or tier objectsLifecycle policiesAlign transitions with access pattern
Prevent deletion/tamperingS3 Object Lock where configuredUnderstand governance/compliance retention behavior at concept level
Query object metadata/inventoryS3 Inventory/AthenaUseful for large-scale reporting
Protect origin contentCloudFront origin access control/origin access identity patternAvoid public bucket origins when private delivery is required
Notes and examples

Block, file, and shared storage

ServiceChoose when…Operations focus
EBSBlock storage for one EC2 instance or supported clustered use caseSnapshots, encryption, volume type/performance, attachment, resizing
EFSShared Linux NFS file systemMount targets, security groups, access points, lifecycle policies
FSx for Windows File ServerManaged Windows SMB file sharesAD integration, backups, Windows workloads
FSx for LustreHigh-performance file system for compute workloadsS3 integration patterns, throughput-heavy jobs
Instance storeTemporary high-performance local storageData is ephemeral; do not use for durable state
S3Object storageEvent notifications, lifecycle, replication, access policies

Database operations

NeedRDS/Aurora featureKey distinction
High availability in a RegionMulti-AZ deploymentHA/failover, not read scaling by itself
Read scalingRead replicas/Aurora replicasCan also support some DR patterns
Point-in-time restoreAutomated backupsRestore creates a new DB resource
Manual long-term recovery pointDB snapshotOperationally controlled backup point
Reduce connection stormsRDS ProxyEspecially useful with spiky/serverless app connections
Diagnose DB loadPerformance Insights, Enhanced Monitoring, CloudWatchChoose based on query/database vs OS-level view
Change engine settingsParameter groupSome changes require reboot depending setting
Upgrade safelySnapshot, test, maintenance window, blue/green where availableAvoid untested production upgrades
Encrypt databaseKMS-backed encryption at creation/restore as supportedPlan key permissions and snapshot sharing behavior
NeedDynamoDB featureKey distinction
Automatic capacity adjustmentAuto scaling or on-demand capacity modeChoose based on predictability
Recover table to prior timePoint-in-time recoveryOperational recovery, not analytics
Global low-latency writes/readsGlobal tablesMulti-Region active-active pattern
React to item changesDynamoDB StreamsFeed Lambda/consumers
Expire old itemsTTLDeletion is asynchronous
Protect accidental deletionBackups, PITR, IAM controlsCloudFormation deletion policy may also matter

S3 operational review

FeatureUse it forTrap
VersioningRecover from overwrite/delete scenariosDelete markers can make objects appear deleted.
Lifecycle policiesTransition or expire objectsUnderstand access pattern before moving to archival classes.
ReplicationSame-Region or cross-Region object replicationVersioning is required; existing objects need special handling.
S3 Block Public AccessPrevent public exposureCan override bucket policies or ACL-based public access.
Object OwnershipControl ownership and ACL behaviorBucket owner enforced disables ACLs.
Object LockWORM retention patternsGovernance and compliance modes differ.
Event notificationsTrigger processing from object eventsWatch destination permissions and event-loop risks.
Storage LensOrganization-wide storage visibilityUseful for usage, activity, and optimization insights.
Access logs / CloudTrail data eventsAudit S3 accessData events can be high volume; scope carefully.

EBS, EFS, and FSx

Storage typeBest fitKey operational point
EBSBlock storage for EC2AZ-scoped; use snapshots for backup and migration.
EFSShared NFS file systemRegional service with mount targets in subnets/AZs.
FSx for Windows File ServerManaged Windows file sharesSMB and Windows integration.
FSx for LustreHigh-performance file system for compute workloadsOften paired with HPC or data processing.
Instance storeTemporary local storageFast but ephemeral.

RDS and database operations

FeaturePurposeExam distinction
Automated backupsPoint-in-time recovery within retentionOperational recovery feature, not read scaling.
Manual snapshotsUser-initiated backups retained until deletedUseful before risky changes.
Multi-AZHigh availability / failoverNot the same as read scaling for traditional RDS deployments.
Read replicasRead scaling and some DR patternsPromotion is a separate action.
Performance InsightsDatabase performance analysisHelps identify waits, SQL load, and bottlenecks.
Enhanced MonitoringOS-level DB instance metricsMore granular than standard CloudWatch DB metrics.
Parameter groupsEngine configurationStatic parameters may require reboot.
Option groupsEngine-specific featuresCommon in certain RDS engines.

Networking and content delivery

VPC connectivity decision table

NeedChooseHigh-yield notes
Public IPv4 internet access for instancePublic subnet route to internet gateway + public IPSecurity group/NACL must allow traffic
Private subnet outbound IPv4 internetNAT gateway or NAT instanceNAT does not allow unsolicited inbound from internet
Private IPv6 outbound internetEgress-only internet gatewayIPv6 does not use NAT in the same way
Private access to S3/DynamoDBGateway VPC endpointRoute table association and endpoint policy matter
Private access to many AWS servicesInterface VPC endpointENI-based, security groups, private DNS option
Connect VPCs at scaleTransit GatewayHub-and-spoke routing; route tables still matter
Simple direct VPC-to-VPC connectivityVPC peeringNon-transitive; CIDR overlap is a blocker
Hybrid encrypted connectionSite-to-Site VPNFaster to establish than physical private connectivity
Dedicated private networkAWS Direct ConnectOften paired with VPN for encryption/backup design
DNS routing and failoverRoute 53Health checks and routing policies are central
Global static entry and accelerationAWS Global AcceleratorRoutes to healthy regional endpoints over AWS network
Cache static/dynamic content at edgeCloudFrontCache behavior, origin, TTL, invalidation, TLS
Notes and examples

Security group vs NACL

ControlSecurity groupNetwork ACL
ScopeElastic network interface/resourceSubnet
StateStatefulStateless
RulesAllow rules onlyAllow and deny rules
EvaluationAll applicable rulesOrdered rule evaluation
Common useInstance/app firewallSubnet guardrail or explicit deny
Exam trapReturn traffic automatically allowedReturn traffic must be explicitly allowed

Route 53 routing policies

PolicyUse when…
SimpleSingle basic answer
WeightedSplit traffic by assigned proportions
Latency-basedSend users to lowest-latency Region
FailoverActive/passive with health checks
GeolocationRoute by user geographic location
GeoproximityRoute by location with optional bias
Multivalue answerReturn multiple healthy records
AliasPoint DNS to supported AWS resources without hard-coding IPs

VPC troubleshooting quick path

    flowchart TD
	    A[Connectivity failure] --> B{DNS resolves?}
	    B -- No --> C[Check Route 53/private hosted zone/resolver/DHCP options]
	    B -- Yes --> D{Route exists?}
	    D -- No --> E[Check route table, TGW, peering, IGW, NAT, endpoint]
	    D -- Yes --> F{Firewall allows?}
	    F -- No --> G[Check security groups and NACLs both directions]
	    F -- Yes --> H{Target healthy/listening?}
	    H -- No --> I[Check OS firewall, app port, ELB health check, instance status]
	    H -- Yes --> J[Check asymmetric routing, TLS, proxy, endpoint policy, IAM]

Security, identity, and compliance operations

IAM policy evaluation reference

ConceptExam meaning
Default denyNo permission unless allowed
Explicit denyOverrides any allow
Identity-based policyAttached to users, groups, or roles
Resource-based policyAttached to resource, such as S3 bucket, KMS key, Lambda function
Permissions boundaryMaximum permissions an identity can receive
SCPMaximum permissions for accounts/OUs in AWS Organizations; does not grant access
Session policyFurther restricts temporary session permissions
RoleAssumed for temporary credentials; preferred for AWS services and cross-account access
Instance profileDelivers IAM role credentials to EC2
Trust policyDefines who can assume a role
Access AnalyzerDetects unintended external access and validates policies
Notes and examples
aws iam simulate-principal-policy \
  --policy-source-arn arn:aws:iam::111122223333:role/AppRole \
  --action-names s3:GetObject \
  --resource-arns arn:aws:s3:::example-bucket/example-key

Security and governance service selection

NeedServiceNotes
API audit logsCloudTrailEnable organization trail for multi-account visibility where appropriate
Resource complianceAWS ConfigManaged/custom rules, aggregators, conformance packs
Threat detectionGuardDutyFindings from account/workload signals
Security findings aggregationSecurity HubConsolidates findings and standards checks
Vulnerability scanningAmazon InspectorEC2, container, and Lambda vulnerability coverage depending configuration
S3 sensitive data discoveryMacieData classification focus
DDoS protectionAWS ShieldStandard is automatic; Advanced adds more protections/features
Web app filteringAWS WAFRules for HTTP/S traffic at ALB, CloudFront, API Gateway, etc.
Certificate managementAWS Certificate ManagerPublic/private cert lifecycle for integrated services
Secrets rotationSecrets ManagerRotation workflows and database integrations
Central account guardrailsAWS Organizations SCPsGuardrails only; IAM still grants actual permissions
Key managementKMSKey policy, IAM, grants, rotation settings, auditing

KMS operational distinctions

TopicRemember
Key policyPrimary control for KMS key access
IAM policyCan allow KMS actions only if key policy permits/delegates
GrantsCommon for AWS services needing temporary/delegated key use
AWS managed keyManaged by AWS for a service/account
Customer managed keyMore control over policy, rotation settings, auditing, deletion scheduling
Multi-Region keyUseful for client-side or service patterns needing related keys across Regions
Encryption contextAdditional authenticated data used by some integrations/policies
S3 SSE-KMS failureCheck both S3 permission and KMS key permission

IAM policy evaluation shortcuts

For SOA-C03 scenarios, remember:

  1. Explicit deny wins.
  2. If there is no applicable allow, the result is implicit deny.
  3. Identity policies, resource policies, permissions boundaries, session policies, and SCPs can all affect the final result.
  4. SCPs do not grant permissions. They set maximum available permissions for accounts or organizational units.
  5. Permission boundaries do not grant permissions. They cap what an identity can do.
  6. Resource policies can allow cross-account access, but the trusted principal may still need identity-side permission depending on the service and action.
  7. Use roles for AWS services and temporary access. Avoid long-term access keys where a role is possible.

IAM scenario table

ScenarioPreferred answer pattern
EC2 instance needs S3 accessAttach an IAM role through an instance profile; do not store access keys on the instance.
Lambda needs to call DynamoDBAdd permissions to the Lambda execution role.
External account needs access to a bucketUse a bucket policy and/or cross-account role with least privilege.
Team needs temporary elevated accessUse IAM Identity Center or role assumption with controlled permissions and audit logging.
Need to limit maximum permissions for developersUse permissions boundaries and appropriate identity policies.
Need guardrails across accountsUse AWS Organizations SCPs.
Need to detect unintended public or cross-account accessUse IAM Access Analyzer.

KMS and secrets

NeedBetter fitKey exam note
Encrypt data with customer-managed access controlAWS KMS customer managed keyKey policy matters; IAM permission alone may not be enough.
Encrypt S3 objectsSSE-S3, SSE-KMS, or client-side encryptionSSE-KMS adds KMS permissions and possible request-cost considerations.
Store database passwords with rotationAWS Secrets ManagerBuilt-in rotation support is a common differentiator.
Store configuration valuesSystems Manager Parameter StoreSecureString can use KMS; advanced features vary by parameter tier.
Temporary credentialsAWS STSPrefer temporary credentials over long-term access keys.

Security monitoring services

ServiceUse it for
Amazon GuardDutyThreat detection from logs and signals such as CloudTrail, VPC flow logs, DNS logs, and more.
AWS Security HubAggregated security findings and posture management.
Amazon InspectorVulnerability management for supported workloads.
Amazon MacieSensitive data discovery for S3.
AWS WAFLayer 7 web request filtering.
AWS ShieldDDoS protection.
IAM Access AnalyzerExternal access and policy validation analysis.

Reliability, backup, and disaster recovery

Reliability design choices

RequirementPreferWhy
Survive instance failureAuto Scaling across Availability ZonesReplaces unhealthy capacity
Survive AZ failure for app tierMulti-AZ subnets + load balancer + Auto ScalingDistributes traffic and capacity
Survive DB instance/AZ failureRDS Multi-AZ or Aurora HA designManaged failover capability
Recover accidental deleteBackups, snapshots, versioning, PITRHA is not backup
Regional disaster recoveryCross-Region backups/replication + Route 53/Global Accelerator patternsDR requires runbooks and testing
Reduce noisy alertsComposite alarms and dependency-aware runbooksAvoid paging on downstream symptoms only
Validate resilienceGame days/failure testing where appropriateKnow rollback and blast radius
Standardize recoverySystems Manager Automation runbooksRepeatable operations beat manual steps
Notes and examples

DR pattern reference

PatternCost/complexityOperational idea
Backup and restoreLowestRestore from backups when needed
Pilot lightLow/mediumCore components replicated; scale out during event
Warm standbyMedium/highScaled-down full environment already running
Active-activeHighestMultiple Regions actively serve traffic

Backup decision points

NeedUse
Centralized policy across supported servicesAWS Backup plans
EC2 volume recoveryEBS snapshots or AWS Backup
RDS database recoveryAutomated backups, snapshots, AWS Backup
S3 object recoveryVersioning, replication, Object Lock where configured
Cross-account backup isolationAWS Backup cross-account strategy
Cross-Region recoveryCross-Region backup/copy/replication
Accidental stack deletion protectionCloudFormation termination protection and deletion policies

Cost, performance, and operational hygiene

GoalTools/actionsExam note
Detect budget overrunAWS Budgets, Cost Explorer, cost anomaly detectionBudgets notify/control; Cost Explorer analyzes
Rightsize computeAWS Compute Optimizer, CloudWatch metricsNeeds enough metric history to make useful recommendations
Reduce idle resourcesFind unattached EBS volumes, idle load balancers, old snapshots, unused Elastic IPsTag ownership and lifecycle
Optimize S3 costLifecycle policies, storage class analysis, inventoryMatch storage class to access and retrieval needs
Reduce NAT dependencyVPC endpoints for supported AWS servicesOften improves private connectivity posture too
Control log costLog retention, filters, export/archive strategyInfinite retention can become expensive
Standardize tagsTag policies, Config rules, cost allocation tagsTags support cost, automation, and ownership
Improve app latencyALB/NLB choice, CloudFront, caching, database tuningDo not solve all latency with larger instances
Improve database performancePerformance Insights, read replicas, indexes/query tuning, cachingMulti-AZ is HA, not a read-scaling feature
Scale queues/workersSQS metrics + Auto Scaling/custom metricsScale on backlog per worker or latency-oriented metric

High-yield traps and distinctions

TrapCorrect exam reasoning
“Need to know who changed it” -> choose CloudWatchChoose CloudTrail for API audit; Config for configuration timeline
“Need to keep resource compliant” -> use CloudTrail onlyUse AWS Config rules or Systems Manager State Manager depending resource/config
“Private subnet needs S3 access” -> NAT is always bestGateway VPC endpoint is usually the private AWS-native path for S3
“Multi-AZ means backup”Multi-AZ is availability; backups/versioning/PITR handle recovery from bad changes
“Read replica means automatic HA failover for primary”Read replicas are primarily read scaling/DR; Multi-AZ is the HA answer for RDS primary failover
“SCP grants admin access”SCPs set maximum permissions; IAM/resource policies still grant
“Security group blocks with deny rule”Security groups allow only; NACLs can explicitly deny
“User data is configuration management”User data bootstraps; Systems Manager/CloudFormation maintain repeatable operations
“Changing launch template updates running instances”Existing instances remain until replaced, refreshed, or relaunched
“CloudFront replaces Route 53”CloudFront caches/distributes content; Route 53 resolves DNS and routes queries
“CloudWatch default EC2 metrics include memory”Memory/disk typically require CloudWatch Agent/custom metrics
“CloudTrail data events are always automatically logged”Know the distinction between management events and optional data-event logging
“KMS IAM allow is enough”Key policy, IAM policy, grants, and service integration all matter
“Public subnet equals internet reachable”Needs route, public address, firewall rules, and listening service
“NACL statefulness works like security groups”NACLs are stateless; return path rules matter

Rapid scenario drill table

If the scenario says…Fast answer direction
“No SSH allowed, but admins need shell access”Systems Manager Session Manager
“Run a command on all instances tagged Environment=Prod”Systems Manager Run Command
“Ensure a package remains installed”Systems Manager State Manager
“Patch instances during a defined window”Patch Manager + Maintenance Windows
“Preview infrastructure changes before update”CloudFormation change set
“Manual console changes caused drift”CloudFormation drift detection; remediate via template
“Notify on EC2 state changes”EventBridge rule
“Alarm when error rate exceeds threshold”CloudWatch metric/math alarm
“Search last hour of app errors”CloudWatch Logs Insights
“Count log pattern as metric”CloudWatch Logs metric filter
“Who opened port 22?”CloudTrail, then Config for current/history
“Block public S3 buckets across accounts”S3 Block Public Access, Config/SCP guardrails as appropriate
“Analyze whether bucket policy allows outside access”IAM Access Analyzer
“Database failover within Region”RDS Multi-AZ/Aurora HA
“Scale reads from database”Read replicas or cache layer
“Static website/global caching”CloudFront in front of S3 or origin
“Private service access from VPC”VPC endpoint
“Hybrid connection over internet”Site-to-Site VPN
“Central hub for many VPCs”Transit Gateway
“Filter malicious HTTP requests”AWS WAF
“Aggregate security findings”Security Hub
“Detect suspicious AWS account activity”GuardDuty
“Find sensitive data in S3”Macie
“Central backup policy and reporting”AWS Backup
“Cost forecast and historical spend”Cost Explorer
“Alert before budget is exceeded”AWS Budgets

Final review checklist

Before sitting for SOA-C03, be able to:

  • Pick between CloudWatch, CloudTrail, AWS Config, EventBridge, and Systems Manager without hesitation.
  • Troubleshoot EC2, Auto Scaling, ELB, Route 53, and VPC connectivity from symptoms.
  • Explain security group vs NACL, NAT vs VPC endpoint, Multi-AZ vs backup, and SCP vs IAM policy.
  • Choose operational automation: Run Command, State Manager, Patch Manager, Automation, CloudFormation, CodeDeploy, or instance refresh.
  • Map storage/database recovery needs to S3 versioning, EBS snapshots, RDS backups/PITR, DynamoDB PITR, and AWS Backup.
  • Recognize least-privilege, encryption, logging, tagging, and repeatable infrastructure patterns.

Next step: work through timed SOA-C03-style scenarios and force yourself to name the AWS service, the operational reason, and the first troubleshooting check before reading the explanation.

Exam mindset: what SOA-C03 rewards

SOA-C03 is not just a service-name recognition exam. Expect scenarios that ask what a CloudOps engineer should do when something is slow, unreachable, noncompliant, under-monitored, over-permissioned, or expensive.

Think in this order:

  1. Observe — metrics, logs, events, traces, audit records, configuration history.
  2. Diagnose — isolate whether the problem is identity, network, compute, storage, quota, scaling, configuration, or dependency related.
  3. Remediate safely — prefer least-disruptive, automated, reversible, and auditable actions.
  4. Harden — apply least privilege, encryption, backup, patching, monitoring, tagging, and drift controls.
  5. Automate — use repeatable AWS-native mechanisms rather than manual console fixes when possible.

Official exam identity

ItemValue
Vendor / providerAWS
Official exam titleAWS Certified CloudOps Engineer – Associate (SOA-C03)
Official exam codeSOA-C03
Page conceptCheat Sheet for final-stage review and question-bank practice

High-yield service map

Operational needAWS services / features to knowExam decision point
Metrics and alarmsAmazon CloudWatch metrics, alarms, anomaly detection, composite alarmsUse metrics for numeric time-series signals; tune periods, thresholds, evaluation periods, and missing-data behavior.
Log collection and searchCloudWatch Logs, Logs Insights, metric filters, subscription filters, CloudWatch AgentUse Logs Insights for investigation; metric filters when log patterns must become alarmable metrics.
Audit API activityAWS CloudTrail, organization trails, data events, management events, Insights eventsCloudTrail answers “who called what API, from where, and when?”
Configuration history and complianceAWS Config, managed rules, custom rules, conformance packs, remediation, aggregatorsConfig answers “what changed, when, and is it compliant?”
Event-driven operationsAmazon EventBridge, EventBridge SchedulerUse events to trigger remediation, workflows, notifications, and operational automation.
Fleet operationsAWS Systems Manager Run Command, Session Manager, Patch Manager, State Manager, Automation, InventoryUse Systems Manager for managed-node operations without SSH/RDP exposure.
Identity and accessIAM users, groups, roles, policies, permission boundaries, STS, resource policies, service-linked rolesExplicit deny wins; roles are preferred for workloads and temporary access.
Network reachabilityVPC route tables, security groups, network ACLs, NAT gateways, internet gateways, VPC endpoints, Transit GatewaySeparate routing, filtering, DNS, and identity problems.
Compute operationsEC2, Auto Scaling groups, launch templates, user data, AMIs, EBS, instance profilesUse immutable images/templates and Auto Scaling health replacement where possible.
Load balancingALB, NLB, Gateway Load Balancer, target groups, health checksMatch protocol layer and target type to the application.
Storage operationsS3, EBS, EFS, FSx, lifecycle policies, replication, snapshotsChoose object, block, or file storage based on access pattern and protocol.
Database operationsRDS, Aurora, DynamoDB, backups, snapshots, Multi-AZ, read replicas, Performance InsightsDistinguish availability, read scaling, backup, and performance diagnosis.
Deployment automationCloudFormation, change sets, drift detection, StackSets, CodeDeploy, CodePipelinePrefer repeatable infrastructure and controlled deployment strategies.
Backup and recoveryAWS Backup, backup plans, vaults, cross-Region / cross-account copyAlign backup design with RPO, RTO, retention, and isolation needs.
Governance and costOrganizations, SCPs, Budgets, Cost Explorer, Cost Anomaly Detection, Trusted Advisor, Compute OptimizerTags, guardrails, and visibility are core operations tools.

Monitoring, logging, and alerting

CloudWatch essentials

FeatureUse it forCommon trap
CloudWatch metricsNumeric signals such as CPU, network, disk, latency, error countMetrics are dimensioned; the wrong dimension can make data appear “missing.”
CloudWatch alarmsNotify or act when metrics breach a conditionAlarms act on state changes, not every repeated datapoint.
Composite alarmsReduce alert noise by combining alarm statesUseful for paging only when multiple symptoms matter together.
Anomaly detectionDynamic baselines for variable workloadsNot a substitute for understanding business-critical thresholds.
CloudWatch LogsCentral log collectionLogs are not the same as CloudTrail audit events.
Logs InsightsInteractive log queryingUse it for investigation, not long-term metric trending by itself.
Metric filtersConvert log patterns into metricsExisting historical log events do not retroactively create metric datapoints.
Subscription filtersStream logs to another destinationUse for near-real-time processing or centralized log pipelines.
CloudWatch AgentOS-level metrics and log collection from EC2/on-premisesDefault EC2 metrics do not include every OS-level metric you may need.
Notes and examples

CloudTrail vs CloudWatch vs Config

Question in the scenarioBest first service
“Who deleted this security group rule?”CloudTrail
“Did CPU or latency spike before the outage?”CloudWatch metrics
“What was the previous configuration?”AWS Config
“Which resources are noncompliant?”AWS Config rules / conformance packs
“Which API calls are unusual?”CloudTrail Insights
“Which log lines contain this error?”CloudWatch Logs Insights
“How do I trigger remediation after an event?”EventBridge + Systems Manager Automation / Lambda

Monitoring traps candidates miss

  • CloudTrail is not a performance monitor. It records API activity; use CloudWatch for metrics and logs.
  • CloudWatch alarms depend on evaluation settings. Period, datapoints to alarm, evaluation periods, and missing-data treatment can change behavior.
  • Not all service metrics are enabled by default at the granularity you want. EC2 detailed monitoring and custom metrics may be required.
  • Logs must be collected before they can be queried. Missing agent configuration or IAM permissions can explain missing logs.
  • A metric filter is not a log search. It creates a metric from matching log events.
  • AWS Config is about resource configuration and compliance, not application traces.
  • Centralized logging often needs cross-account design. Consider organization trails, delegated admin, log archive accounts, and resource policies.

Incident response decision flow

    flowchart TD
	    A[Alarm, ticket, or user report] --> B{Is it availability, performance, security, or compliance?}
	    B -->|Availability / performance| C[Check CloudWatch metrics, alarms, target health, logs]
	    B -->|Security / audit| D[Check CloudTrail, IAM Access Analyzer, GuardDuty, Security Hub]
	    B -->|Compliance / drift| E[Check AWS Config history, rules, conformance packs]
	    C --> F{Likely resource layer?}
	    F -->|Network| G[Routes, security groups, NACLs, DNS, endpoints, load balancer]
	    F -->|Compute| H[EC2 status checks, ASG health, ECS/Lambda errors, capacity]
	    F -->|Storage / database| I[S3/KMS policies, EBS/EFS, RDS metrics, backups, replicas]
	    D --> J[Contain, revoke, rotate, patch, document]
	    E --> K[Remediate with SSM Automation, CloudFormation, or Config remediation]
	    G --> L[Apply least disruptive fix]
	    H --> L
	    I --> L
	    J --> M[Post-incident: add alarms, rules, runbooks, and tests]
	    K --> M
	    L --> M

Networking and connectivity

VPC reachability checklist

When a resource cannot connect, separate the layers:

LayerWhat to check
DNSCorrect hostname, private hosted zone association, resolver rules, split-horizon behavior.
RouteRoute table has a matching route to internet gateway, NAT gateway, VPC peering, Transit Gateway, VPN, or VPC endpoint.
Source / destinationPublic subnet workloads need public IP or Elastic IP for direct internet access.
Security groupStateful allow rules on the ENI or attached resource.
Network ACLStateless inbound and outbound rules; ephemeral ports matter.
Endpoint policyVPC endpoint policy may block access even if IAM allows it.
Resource policyS3 bucket policy, KMS key policy, SQS queue policy, etc.
OS / applicationLocal firewall, listener port, service health, certificate, proxy, or application config.
Notes and examples

Security groups vs network ACLs

FeatureSecurity groupNetwork ACL
ScopeENI / resource levelSubnet level
StateStatefulStateless
RulesAllow rules onlyAllow and deny rules
EvaluationAll rules consideredNumbered rules evaluated in order
Common useInstance/application access controlSubnet-level guardrail or explicit deny pattern
Exam trap“Outbound response traffic” is automatically allowed for established flowsMust allow return traffic, often including ephemeral ports

Internet, private access, and hybrid access

RequirementCommon AWS pattern
Public web appPublic subnet ALB + private subnet targets where possible
Private subnet instance needs outbound internet updatesNAT gateway in a public subnet + route from private subnet
Private access to S3 or DynamoDBGateway VPC endpoint
Private access to many AWS servicesInterface VPC endpoint powered by AWS PrivateLink
Connect VPCs at scaleAWS Transit Gateway
Simple non-transitive VPC-to-VPC connectionVPC peering
Encrypted internet-based hybrid connectionAWS Site-to-Site VPN
Dedicated private hybrid connectivityAWS Direct Connect
Centralized egress inspectionTransit Gateway + inspection VPC / network firewall pattern

Route 53 review

Routing policyUse when
SimpleOne basic answer for a name.
WeightedShift a percentage of traffic between targets.
Latency-basedSend users to the lowest-latency AWS Region.
FailoverActive-passive DNS failover using health checks.
GeolocationRoute based on user location.
GeoproximityRoute based on location and optional bias.
Multivalue answerReturn multiple healthy records.

Infrastructure as Code and configuration control

CloudFormation review

FeatureWhat to remember
StacksDeploy and manage related AWS resources as a unit.
Change setsPreview proposed changes before execution.
Drift detectionIdentify resources changed outside CloudFormation.
Stack policiesProtect critical stack resources from unintended updates.
DeletionPolicyRetain, snapshot, or delete selected resources on stack deletion.
RollbackFailed updates can roll back to a previous known state.
StackSetsDeploy stacks across accounts and Regions.
Parameters / mappings / conditionsReuse templates across environments.
Outputs / exportsShare values with other stacks.
Notes and examples

CloudFormation traps

  • Drift detection tells you drift exists; it does not automatically fix every issue.
  • Change sets are previews, not guarantees of successful deployment.
  • Manual console changes create operational risk when resources are supposed to be managed by IaC.
  • Stack deletion can delete resources unless protected with appropriate policies.
  • StackSets are for multi-account / multi-Region rollout, not for single-resource troubleshooting.

Backup, recovery, and resilience

RTO and RPO

TermMeaningDesign implication
RTOHow quickly service must be restoredDrives standby architecture and automation.
RPOHow much data loss is acceptableDrives backup frequency, replication, and durability design.
Notes and examples

Recovery patterns

RequirementLikely pattern
Recover from accidental file/object deletionVersioning, snapshots, backups, restore testing.
Recover EC2 workload quicklyAMI, launch template, Auto Scaling, EBS snapshots.
Protect against AZ failureMulti-AZ design, load balancing, Auto Scaling across AZs.
Protect against Region-level issueCross-Region replication, backups, Route 53 failover, tested runbooks.
Centralized backup policyAWS Backup plans and vaults.
Isolated backup copiesCross-account and/or cross-Region copy.
Database high availabilityRDS/Aurora Multi-AZ patterns.
Static website/object recoveryS3 versioning, replication, lifecycle, backup strategy.

Resilience traps

  • Backups are only useful if restores are tested.
  • Multi-AZ improves availability; it is not the same as multi-Region disaster recovery.
  • Read replicas may improve read performance but do not automatically solve every failover requirement.
  • DNS failover depends on health checks, TTLs, and application readiness.
  • Single NAT gateways, single-AZ databases, and single load balancer target groups can hide availability risks.

Containers and serverless operations

Lambda operations

TopicReview point
TimeoutLong-running functions fail if timeout is too low.
MemoryAlso affects CPU allocation; increasing memory can improve performance.
ConcurrencyThrottling can occur at account or function concurrency limits.
Reserved concurrencyGuarantees and caps concurrency for a function.
Provisioned concurrencyReduces cold-start impact for predictable workloads.
DLQ / destinationsHandle asynchronous invocation failures.
Environment variablesConfiguration; use KMS or secrets services for sensitive values.
VPC accessNeeded for private resources but can introduce networking considerations.
CloudWatch LogsPrimary place for function logs.
Notes and examples

ECS operational points

NeedECS feature
Define containersTask definition
Run and maintain desired countECS service
Serverless container computeFargate
EC2-backed container capacityECS on EC2 with capacity providers
Service discovery / load balancingCloud Map, ALB/NLB integration
Logsawslogs driver to CloudWatch Logs
SecretsSecrets Manager or Parameter Store integration
Deployment safetyRolling updates, circuit breaker, blue/green with CodeDeploy

Cost, governance, and operational excellence

Cost and optimization tools

ToolUse it for
AWS Cost ExplorerAnalyze historical and forecasted cost usage.
AWS BudgetsAlert on cost, usage, reservation, or savings-plan thresholds.
Cost Anomaly DetectionDetect unusual spend patterns.
Cost and Usage ReportDetailed billing data for analysis.
Trusted AdvisorRecommendations across cost, security, fault tolerance, performance, and service limits, depending on support plan.
Compute OptimizerRightsizing recommendations for supported compute resources.
S3 Storage LensStorage usage and optimization visibility.
Cost allocation tagsAttribute spend to teams, apps, or environments.
Notes and examples

Governance review

RequirementAWS pattern
Separate production, staging, and devMulti-account strategy with AWS Organizations.
Prevent disallowed services or RegionsSCP guardrails.
Standardize account baselinesControl Tower / account vending patterns.
Centralize audit logsOrganization CloudTrail and log archive account.
Detect resource driftAWS Config and CloudFormation drift detection.
Enforce taggingTag policies, Config rules, IaC validation, automation.
Centralize security findingsSecurity Hub delegated administration.

Common troubleshooting patterns

EC2 instance is unreachable

Check in this order:

  1. Instance state and status checks.
  2. Correct public/private IP and DNS name.
  3. Route table path.
  4. Security group inbound rule.
  5. Network ACL inbound and outbound rules.
  6. OS firewall and service listener.
  7. Key pair / login method / SSM Session Manager availability.
  8. IAM role and SSM Agent if using Session Manager.
  9. Recent CloudTrail or Config changes.

Application behind ALB returns 5xx

SymptomLikely check
No healthy targetsTarget group health check path, port, protocol, security group.
Intermittent errors during deployDeployment strategy, deregistration delay, readiness checks.
High latencyTarget CPU/memory, database dependency, scaling policy, connection behavior.
TLS issueListener certificate, security policy, target protocol.
Wrong routingListener rules, host/path conditions, priority order.
Notes and examples

S3 AccessDenied

Review:

  • IAM identity policy.
  • Bucket policy.
  • S3 Block Public Access.
  • Object ownership and ACL behavior.
  • KMS key policy and KMS permissions for SSE-KMS objects.
  • VPC endpoint policy if access is through an endpoint.
  • SCPs or permission boundaries.
  • Object key name, prefix condition, encryption condition, or source IP/VPC condition.

RDS connection issue

CheckWhy it matters
DB instance statusInstance may be modifying, backing up, failing over, or unavailable.
Security groupClient source must be allowed to DB port.
Subnet and routingPrivate DBs require network path from clients.
Public accessibilityPublic flag alone is not enough; routing and SGs still matter.
DNS endpointEndpoint may change after failover.
Parameter groupConnection limits or SSL settings may affect access.
CPU, memory, storageResource exhaustion causes timeouts and failures.
Logs / Performance InsightsIdentify engine-level errors and waits.

Lambda failures

SymptomCommon cause
TimeoutFunction timeout too low, dependency slow, VPC/network issue.
ThrottlingConcurrency limit or reserved concurrency setting.
Access deniedExecution role lacks permission or KMS/resource policy blocks access.
No logsRole lacks logging permission or function did not initialize.
Async retries exhaustedConfigure DLQ or destination and inspect failure payloads.
Cannot reach private resourceVPC config, route table, security group, DNS, or endpoint issue.

“Best answer” decision rules

Use these fast rules when two answers seem plausible:

If the scenario says…Prefer…
“Without opening inbound SSH/RDP”Systems Manager Session Manager / Run Command
“Who made this change?”CloudTrail
“What changed in resource configuration?”AWS Config
“Search application logs”CloudWatch Logs Insights
“Turn matching logs into an alarm”CloudWatch Logs metric filter + CloudWatch alarm
“Private subnet needs internet updates”NAT gateway route
“Private access to S3”Gateway VPC endpoint
“Private access to AWS APIs”Interface VPC endpoint
“Cross-account guardrails”AWS Organizations SCPs
“Temporary access”IAM role / STS
“EC2 app needs AWS API access”Instance profile role
“Store and rotate DB credentials”Secrets Manager
“Repeatable infrastructure deployment”CloudFormation
“Preview infrastructure changes”CloudFormation change set
“Detect infrastructure drift”CloudFormation drift detection or AWS Config, depending on scope
“Patch a fleet on a schedule”Systems Manager Patch Manager + Maintenance Windows
“Automated operational runbook”Systems Manager Automation
“Scale based on utilization target”Auto Scaling target tracking
“HTTP path-based routing”ALB
“TCP/UDP high-performance load balancing”NLB
“Recover deleted S3 objects”Versioning / backups
“Database HA failover”Multi-AZ
“Read scaling”Read replicas
“Central backup policy”AWS Backup

Practice priorities for SOA-C03

Use the Cheat Sheet as a checklist, then validate with original practice questions. Prioritize topic drills in this order if your time is limited:

  1. Monitoring and incident response — CloudWatch, CloudTrail, Config, EventBridge, Systems Manager.
  2. Networking troubleshooting — VPC routes, security groups, NACLs, endpoints, NAT, load balancers, Route 53.
  3. IAM and security operations — policy evaluation, roles, KMS, secrets, audit, detective controls.
  4. Compute and scaling — EC2 status checks, Auto Scaling, ALB/NLB, deployment patterns.
  5. Storage and databases — S3 access, lifecycle, replication, EBS/EFS, RDS backups, Multi-AZ, read replicas.
  6. Automation and IaC — CloudFormation, drift, change sets, StackSets, Systems Manager Automation.
  7. Backup, resilience, and cost governance — AWS Backup, tagging, Budgets, Cost Explorer, Organizations.

How to review explanations effectively

After each question-bank item, ask:

  • Which AWS service is the scenario really testing?
  • Is the issue about identity, networking, configuration, capacity, availability, or observability?
  • Which option is the most operationally safe and least manual?
  • Which option violates least privilege, high availability, or automation principles?
  • Did the wrong answer solve a similar problem but not the exact requirement?
  • Did the question ask for prevention, detection, remediation, or investigation?

Final quick-check list

Before your next mock exam, make sure you can explain:

  • CloudWatch vs CloudTrail vs AWS Config.
  • Security groups vs network ACLs.
  • NAT gateway vs internet gateway vs VPC endpoint.
  • IAM role vs user vs resource policy vs SCP.
  • Secrets Manager vs Parameter Store.
  • ALB vs NLB vs Gateway Load Balancer.
  • Auto Scaling target tracking vs step vs scheduled scaling.
  • RDS Multi-AZ vs read replica.
  • S3 versioning, lifecycle, replication, Block Public Access, and KMS access.
  • Systems Manager Session Manager, Run Command, Patch Manager, State Manager, and Automation.
  • CloudFormation change sets, drift detection, stack policies, deletion policies, and StackSets.
  • Backup design using RTO, RPO, retention, and restore testing.

Put the review into practice