AWS Config records supported resource settings and relationships and evaluates them against compliance rules. Systems Manager Automation runbooks perform configured fixes, organization conformance packs distribute checks and remediation, and aggregators collect results into a central read-only view.
Recorders and configuration history
AWS Config records the settings and relationships of supported AWS resources, entities such as instances and storage volumes, so you can investigate how they changed. A configuration itemA point-in-time record of a supported resource's configuration, attributes, and relationships. (CI) is a record of one resourceAn entity created and managed in AWS, such as an EC2 instance, storage volume, or security group. at a particular time. Its configuration historyThe sequence of recorded configuration items showing how a resource changed over time. starts when recording starts.
How it works
-
Choose resource types in each account and Region. The customer managed recorderThe Config component that stores configuration changes for the resource types selected for recording., whose scopeThe resource types, identifiers, or tags used to decide which resource changes initiate a rule evaluation. you control, defaults to supported types except global IAM users, groups, roles, and customer managed policies. A service-linked recorderA recorder whose resource scope is set by its linked AWS service. has its scope set by its linked service.
-
Select ContinuousRecording configuration changes whenever they occur. recording for changes as they occur or DailyRecording the latest resource state over a 24-hour period only when it differs from the previous recorded configuration item. recording for the latest state over 24 hours, only when different from the previous CI. Repeated creation and deletion of temporary resources can increase recording activity.
-
Review a resource’s configuration history, its sequence of CIs, through its console timeline or the API. A configuration snapshotA collection of configuration items for the resources being recorded in an account. collects the recorded resources’ configurations together.
-
Use the delivery channelThe Config destination settings controlling where configuration updates and notifications are sent., the destination settings for configuration updates, to deliver history to S3. History files arrive every six hours for resource types with changes.
DeliverConfigSnapshotrequests a snapshot. Config does not manage the delivered objects’ S3 lifecycle policies, which control archiving and deletion. -
For patch and State Manager association complianceWhether a resource's configuration satisfies the conditions of a rule. history, record
AWS::SSM::PatchComplianceandAWS::SSM::AssociationCompliance. Config also supports Organizations account, organization, organizational unitA group of AWS accounts within an organization that can also contain other organizational units., policy, and resource policy types. An organizational unit groups accounts within an organization. -
Protect the recording baseline with a service control policyAn organization policy limiting the maximum permissions available to IAM users and roles in member accounts. (SCP), a limit on member-account IAM permissions. Deny recorder stop, delete, and update actions plus delivery-channel delete and update actions, while retaining other permitted administration. AWS documents this pattern in its Control Tower Config protection SCP.
Exam cues
- If the question says investigate a resource’s settings before an incident, think configuration history, because its CIs show configurations and relationships over time from the start of recording.
- If the question says a configuration-change rule has no evaluations, think recorder status, selected resource types, and rule scope, because those evaluations follow matching CI change notifications.
- If the question says retain patch and association compliance changes, think Config recording of Systems Manager compliance resources, because current compliance reports alone do not provide that history.
- If the question says Firewall Manager must monitor changes, think Continuous recording, because Firewall Manager depends on it.
Remember
- One customer managed recorder per account per Region.
- To protect the baseline, deny
config:DeleteConfigurationRecorder,config:StopConfigurationRecorder,config:PutConfigurationRecorder,config:DeleteDeliveryChannel, andconfig:PutDeliveryChannelthrough an SCP. - Daily recording represents the latest state over 24 hours, not every intermediate change.
- Global IAM types require Regions where Config was supported before February 2022. Record them once in a supported Region to avoid duplicates.
- Supported global CloudFront, Route 53, WAF, ECR Public, and Global Accelerator types use their designated home RegionThe designated Region in which Config records a particular global resource type., the Region where Config records that global type in the commercial partition. Aurora global clusters are an exception and are recorded in enabled supported Regions unless excluded.
- Non-recorded resources have creation and deletion markers, not a configuration-change history on their own page.
RetentionPeriodInDaysaccepts 30 through 2557 days for CI history; the default is seven years, or 2557 days. S3 lifecycle policies control delivered objects separately.
Trap: A missing CI does not prove a periodicA rule trigger that starts evaluation at a selected frequency. rule cannot run. Periodic rules do not depend on CIs; configuration-change rules do.
The recorder captures the selected resource's configuration and relationships as configuration items.
Config delivers history files to the configured S3 bucket; the console timeline also exposes resource history.
Box details (3)
- EC2 instance
A supported resource selected for recording.
- Config recorder
Records configuration items at the selected recording frequency.
- S3 history
Receives history files for changed resource types every six hours.
From the AWS docs (30)
- A configuration item is a point-in-time record of a supported resource's attributes, configuration, relationships, and related events.
- AWS Config maintains configuration history from the time the configuration recorder starts.
- Configuration history is the sequence of configuration items for a resource, accessible through its console timeline or the API.
- A configuration snapshot collects the configuration items for the resources being recorded in an account.
- AWS Config delivers JSON configuration history files to the specified S3 bucket every six hours for resource types with changes during that period.
- DeliverConfigSnapshot requests delivery of a configuration snapshot to the configured S3 bucket.
- AWS Config does not change the lifecycle policies of the S3 objects it delivers.
- A customer managed recorder lets you include or exclude resource types and defaults to all supported types except global IAM users, groups, roles, and customer managed policies.
- Each account can have one customer managed configuration recorder per Region.
- A service-linked recorder has its resource scope set by its linked AWS service.
- Continuous recording captures configuration changes whenever they occur.
- Daily recording captures the latest state over a 24-hour period only when it differs from the previous recorded configuration item.
- Firewall Manager depends on continuous recording to monitor resources.
- Global IAM users, groups, roles, and customer managed policies can be recorded only in Regions where Config was supported before February 2022; recording them in one supported Region avoids duplicates.
- Config records supported global CloudFront, Route 53, WAF, ECR Public, and Global Accelerator resource types only in their designated home Region in the commercial partition.
- AWS::RDS::GlobalCluster is recorded in all enabled supported Config Regions unless excluded from recording.
- Non-recorded resources have creation and deletion markers but no configuration-change history or relationship details on their own resource page.
- Periodic rules can run without the configuration recorder enabled because they do not depend on configuration items.
- A configuration-change rule evaluates after a configuration item change notification for a resource matching its scope.
- Recording Systems Manager patch and association compliance resources in Config enables compliance history and change tracking.
- Supported Config resource types include AWS::Organizations::Account, AWS::Organizations::Organization, AWS::Organizations::OrganizationalUnit, AWS::Organizations::Policy, and AWS::Organizations::ResourcePolicy.
- RetentionPeriodInDays controls configuration item history retention and accepts 30 through 2557 days.
- Temporary workloads can increase Config recording and evaluation activity through repeated resource creation and deletion.
- The delivery channel controls the destinations for Config notifications and configuration updates.
- Config retains configuration items for seven years, or 2557 days, by default.
- An organizational unit groups AWS accounts within an organization and can contain other organizational units.
- AWS::SSM::PatchCompliance and AWS::SSM::AssociationCompliance are supported Config resource types.
- A service control policy limits the maximum permissions of IAM users and roles in an organization.
- The Control Tower Config protection SCP denies DeleteConfigurationRecorder, StopConfigurationRecorder, PutConfigurationRecorder, DeleteDeliveryChannel, and PutDeliveryChannel, with an exception for AWSControlTowerExecution.
- An AWS resource is an entity you create and manage in AWS, such as an EC2 instance or storage volume.
Managed, custom, and proactive rules
An AWS Config rule compares resourceAn entity created and managed in AWS, such as an EC2 instance, storage volume, or security group. settings with desired conditions and reports complianceWhether a resource's configuration satisfies the conditions of a rule., meaning whether the resource passes the check. Choose the rule’s implementation separately from its trigger, which decides when to run, and its evaluation mode, which decides whether to assess deployed or proposed resources.
How it works
- Pick a managed ruleA predefined Config compliance check created by AWS with customizable settings. and its parameters, or write a custom rule. GuardA policy-as-code language for expressing checks on resource configuration properties. is a language for expressing configuration checks as policies. A custom Lambda rule runs your code and reports results with
PutEvaluations. - Define the rule’s scopeThe resource types, identifiers, or tags used to decide which resource changes initiate a rule evaluation., the resources whose changes initiate evaluations: resource types, a type and resource ID, a tagA label with a key and an optional value used to describe a resource. key and value, or all recorded changes. A tag is a key with an optional value used to label a resource.
- Config evaluates existing resources when a rule is added. Later evaluations follow the rule’s trigger. Detective custom Guard rules use configuration changes.
| Trigger | When evaluation runs |
|---|---|
| Configuration changes | After a CIA point-in-time record of a supported resource's configuration, attributes, and relationships. change notification for a resource matching scope |
| PeriodicA rule trigger that starts evaluation at a selected frequency. | At the selected frequency |
| HybridA rule trigger combining configuration-change evaluations with periodic evaluations. | Both on changes and periodically |
- Use detective evaluationEvaluation of the configuration settings of already deployed resources. for deployed resources and proactive evaluationEvaluation of proposed resource properties before deployment without blocking deployment or applying remediation. for proposed properties before deployment. A pipeline must enforce its own rejection decision.
- To see evaluations as Security Hub CSPM findingsA report describing a security or compliance observation about an affected resource., activate the AWS Config integration. It sends managed and custom evaluations through EventBridge. A finding is a security or compliance report about a resource.
Exam cues
- If the question says find running EC2 instances using unapproved images, think
approved-amis-by-idwithamiIds, because it compares their Amazon Machine ImageAn image supplying the software needed to set up and boot an EC2 instance. (AMI) IDs with the approved list. An AMI supplies the software used to boot an instance. - If the question says check encryption on attached EBS volumes, think
encrypted-volumes, because it checks encryption and can require the KMS key inkmsId. - If the question says detect that no CloudTrail trail is enabled, think
cloudtrail-enabled, because it checks that condition periodically. Evaluation alone does not configure a fix. - If the question says check recorded properties without writing a Lambda function, think a Guard custom policy ruleA Config rule whose configuration checks are expressed in Guard rather than a custom Lambda function., because the policy expresses the check.
- If the question says reject noncompliant proposed infrastructure before deployment, think proactive evaluation plus a pipeline decision, because Config supplies a result without blocking deployment.
- If the question says check S3 public read or write access, think
s3-bucket-public-read-prohibitedors3-bucket-public-write-prohibited. Both consider Block Public Access, the bucket policy, and the bucket access control list (ACL). For public bucket-level settings with exclusions, thinks3-bucket-level-public-access-prohibitedandexcludedPublicBuckets.
Remember
- Trigger type and evaluation mode answer different questions: when to run and which resource state to assess.
- Proactive evaluation does not remediate or prevent deployment.
- The Security Hub CSPM integration sends evaluations performed after Security Hub CSPM was enabled. Service-linked rules are excluded except rules created and managed by Control Tower.
Trap: A rule reports noncompliance; a remediation actionA corrective action that Config runs through a Systems Manager Automation document for a resource reported as noncompliant. performs the fix. A proactive result is not a deployment block.
Trap:
encrypted-volumeschecks attached volumes. Do not assume it checks every unattached volume.
Enable proactive evaluation on a supporting rule, then submit properties with StartResourceEvaluation.
Use the returned evaluation ID with GetResourceEvaluationSummary. To reject a failed check, the pipeline must enforce that decision.
Box details (3)
- Pipeline input
The pipeline submits proposed resource properties for a supported resource type.
- Config rules
StartResourceEvaluation starts the check. Config evaluates proactive rules in this account and Region.
- Pipeline gate
The pipeline retrieves the summary and enforces its decision; Config does not prevent deployment.
From the AWS docs (27)
- Config rules evaluate resource configuration settings against desired conditions and flag failures as noncompliant.
- Managed rules are predefined customizable rules created by AWS Config.
- Custom rules use either Lambda functions for custom code or Guard policies for policy-as-code checks without developing a Lambda function.
- Config invokes a custom rule's Lambda function on configuration changes or periodically; the function sends results using PutEvaluations.
- Detective evaluations for custom Guard policy rules are initiated by configuration changes.
- After a rule is added, Config compares existing resources with its conditions before subsequent triggered evaluations.
- Configuration-change evaluations follow a CI change notification for a resource matching the rule scope.
- Rule scope can select resource types, a resource type with an ID, a tag key and value, or all recorded resource changes.
- Periodic rules run at a selected frequency.
- Hybrid rules evaluate both on configuration changes and periodically.
- Detective evaluation checks the configurations of already deployed resources.
- Proactive evaluation checks proposed resource properties against proactive rules in the account and Region before deployment.
- Proactive rules neither prevent deployment nor remediate resources flagged as NON_COMPLIANT.
- After enabling PROACTIVE evaluation on a supporting rule, call StartResourceEvaluation with resource properties and use its ResourceEvaluationId with GetResourceEvaluationSummary to retrieve the result.
- The configuration-change rule approved-amis-by-id checks EC2 instances against the amiIds list and marks running instances with unlisted AMIs NON_COMPLIANT.
- The configuration-change rule encrypted-volumes checks attached EBS volumes for encryption and optionally checks the KMS key specified by kmsId.
- The periodic cloudtrail-enabled rule reports NON_COMPLIANT if a CloudTrail trail is not enabled.
- The s3-bucket-public-read-prohibited rule evaluates public read access using S3 Block Public Access settings, the bucket policy, and bucket ACL.
- The s3-bucket-public-write-prohibited rule evaluates public write access using S3 Block Public Access settings, the bucket policy, and bucket ACL.
- The AWS Config rule `s3-bucket-level-public-access-prohibited` flags public bucket-level settings unless the bucket is listed in excludedPublicBuckets.
- The Config integration sends managed and custom rule evaluations through EventBridge as Security Hub CSPM findings after the integration is activated.
- The Config integration sends evaluations performed after Security Hub CSPM is enabled and excludes service-linked rules except those created and managed by Control Tower.
- A tag is a resource label consisting of a key and an optional value.
- An Amazon Machine Image provides the software needed to set up and boot an EC2 instance.
- EC2 tenancy choices distinguish shared instances, Dedicated Instances, and Dedicated Hosts.
- A Dedicated Host is a physical server with EC2 instance capacity dedicated to your use.
- Security Hub CSPM findings contain a security observation and identify the affected resource and current status.
Automation remediation and retries
AWS Config remediationA corrective action that Config runs through a Systems Manager Automation document for a resource reported as noncompliant. corrects a resourceAn entity created and managed in AWS, such as an EC2 instance, storage volume, or security group. that a rule reports as noncompliantWhether a resource's configuration satisfies the conditions of a rule.. The rule detects the problem; a Systems Manager Automation runbookA document defining the actions Systems Manager performs as a sequence of steps., a document listing corrective actions, performs the fix.
How it works
- Associate a managed or custom Automation runbook with a supported Config rule. For rules created by a conformance packA template-defined bundle of Config rules and remediation actions deployed together., put the remediation in the pack template and update the pack. Direct
PutRemediationConfigurationsdoes not support service-linked rules, including organization rules. - Supply the runbook’s parameters. A static parameterA remediation parameter with a fixed value rather than a resource ID substituted at runtime. stays fixed; a dynamic parameterA remediation parameter whose value is replaced with the target resource ID at runtime. receives the noncompliant resource ID at runtime.
- Choose manual or automatic remediation. Manual waits for an operator to select resources and start the action. Automatic starts for reported noncompliance and requires
AutomationAssumeRole, the IAM roleAn assumable identity whose permissions determine which AWS actions it can perform. whose permissions Automation uses. - Configure retries for failed automatic actions within a time window.
- For notifications, match
aws.configevents of typeConfig Rules Compliance Changein EventBridge and target an SNS topicA communication channel used to publish messages to subscribing endpoints., a channel for publishing messages to subscribers.ComplianceChangeNotificationidentifies a changed compliance result. This notification path is separate from automatic remediation.
Exam cues
- If the question says automatically correct noncompliant resources, think Config remediation plus an Automation runbook, because the evaluation and corrective action have separate responsibilities.
- If the question says alert on unrestricted incoming SSH but let a person decide the fix, think
restricted-ssh, EventBridge, SNS, and manual remediation, because automatic correction is not required for notification. - If the question says detect old active IAM user access keys, think periodicA rule trigger that starts evaluation at a selected frequency.
access-keys-rotatedwithmaxAccessKeyAge, because it compares key age with the configured threshold. - If the question says deploy checks and fixes across member accounts, think an organization conformance pack, because it packages both in the deployment template.
- If the question says enable S3 server access logging for buckets without it, think
s3-bucket-logging-enabledandAWS-ConfigureS3BucketLogging, because the rule detects missing logging and the runbook can remediate it automatically.
Remember
-
Recording covers selected supported resource types. Continuous recordingRecording configuration changes whenever they occur. captures changes as they happen; daily recordingRecording the latest resource state over a 24-hour period only when it differs from the previous recorded configuration item. retains only the latest changed state over 24 hours.
-
access-keys-rotateddefaults to 90 days and excludes root user keys. -
restricted-sshchecks unrestricted SSH from0.0.0.0/0or::/0, meaning any IPv4 or IPv6 address. It has both change and periodic triggers. -
Automatic remediation requires
AutomationAssumeRoleeven when a runbook can otherwise run without an explicit role. -
Config event delivery to EventBridge is best effort.
Trap: An EventBridge notification does not automatically fix the resource. Automatic remediation is configured separately.
A configuration change to a selected resource produces a configuration item at the recording frequency.
A configuration-change rule evaluates matching changes; periodic evaluations run on their schedule.
Configured automatic remediation starts a runbook for reported noncompliance; manual remediation waits for an operator.
An aggregator independently collects source configuration and compliance data into a central read-only view.
Compliance changes can independently trigger EventBridge notifications or other workflows.
Box details (6)
- EC2 instance
An example supported resource selected for recording.
- Config recorder
Records selected resource configurations. Periodic rule evaluations do not depend on configuration items.
- Config rule
Evaluates on matching configuration changes, a periodic schedule, or both, depending on the rule.
- SSM Automation
Automatic remediation uses a runbook, parameters, and AutomationAssumeRole. Manual remediation waits for an operator.
- Aggregator
Reads replicated configuration and compliance data. It cannot deploy rules into source accounts.
- EventBridge
Matches compliance change events for notifications or other workflows. Config event delivery is best effort.
From the AWS docs (25)
- Config remediation runs a managed or custom Systems Manager Automation document defining the corrective actions.
- Manual remediation requires an operator to select resources and start the configured remediation action.
- Automatic remediation starts configured corrective actions for resources reported as noncompliant.
- Automatic Config remediation requires a value for AutomationAssumeRole.
- A remediation parameter can use a static value or dynamically receive the noncompliant resource ID.
- PutRemediationConfigurations cannot add remediation to service-linked rules, including organization rules, pack rules, and Security Hub rules.
- Add remediation for a service-linked pack rule to the conformance pack template and update the pack.
- MaximumAutomaticAttempts is the maximum failed automatic-remediation attempts, defaults to 5, and accepts 1 through 25.
- RetryAttemptSeconds is the time window for counting attempts before adding a remediation exception and defaults to 60 seconds.
- EventBridge can match Config Rules Compliance Change events from aws.config and ComplianceChangeNotification messages to react to changed compliance.
- Config emits events to EventBridge on a best-effort basis.
- An EventBridge rule can send matching events to an Amazon SNS topic.
- The periodic access-keys-rotated rule checks whether active IAM access keys exceed maxAccessKeyAge, which defaults to 90 days.
- The access-keys-rotated rule does not evaluate root user access keys.
- The restricted-ssh rule flags unrestricted incoming SSH access from 0.0.0.0/0 or ::/0 as NON_COMPLIANT and uses configuration-change and periodic triggers.
- An organization conformance pack deploys Config rules and remediation actions together across member accounts.
- AWS Config records only selected supported resource types; continuous recording captures changes as they occur, while daily recording keeps the latest state in a 24-hour period only if it differs from the previous recorded state.
- Config rule evaluations can follow configuration changes, a periodic schedule, or both.
- An aggregator provides a read-only copy of source account and Region configuration and compliance data.
- An Automation runbook defines the actions that Systems Manager performs in sequential steps.
- An IAM role is an assumable identity with permissions governing the actions it may perform.
- PutRemediationConfigurations lists iam:PassRole to ssm.amazonaws.com as a dependent permission.
- iam:PassRole permits a caller to pass an IAM role to an AWS service for that service to assume later.
- An SNS topic is a communication channel delivering messages to subscribing endpoints.
- The managed rule s3-bucket-logging-enabled flags buckets without server access logging, and the AWS-ConfigureS3BucketLogging runbook can remediate them automatically.
Conformance packs and organization rules
An AWS Config conformance packA template-defined bundle of Config rules and remediation actions deployed together. bundles complianceWhether a resource's configuration satisfies the conditions of a rule. rules and remediation actionsA corrective action that Config runs through a Systems Manager Automation document for a resource reported as noncompliant. into one deployment. Use an organization ruleA Config rule centrally deployed and managed across accounts in an organization. for a centrally managed ruleA predefined Config compliance check created by AWS with customizable settings., or an organization conformance pack when the deployment should include a collection of checks and fixes.
How it works
- Enable Config recording in the target accounts and Regions. For organization packs, enable all Organizations features. Deploy from the management account or a delegated administratorA member account authorized to administer a specific integrated service across an organization., a member account authorized to administer the integration. Register that delegate for
config-multiaccountsetup.amazonaws.com. - Write a YAML templateA text document used to describe the Config rules and remediation actions in a conformance pack., a text file describing the pack’s managed or custom rules and remediation actions. Deploy it with
PutOrganizationConformancePack. For individual organization rules, usePutOrganizationConfigRulethrough the API or CLI; their creation is unavailable in the Config console. - Deploy separately in each target Region. Specify excluded accounts when some members should not receive a pack. Member accounts cannot modify centrally deployed pack rules and remediation actions.
- As accounts join, Config deploys the existing organization rules and packs to them. When accounts leave, Config removes those deployments. If a new account has no recorderThe Config component that stores configuration changes for the resource types selected for recording., deployment retries for seven hours after it joins.
- To add or change remediation for a rule in a pack, update the pack template. These are service-linked rules, rules owned by the service that creates them. Direct
PutRemediationConfigurationscalls do not support them.
Exam cues
- If the question says deploy common checks and fixes to member accounts, think an organization conformance pack, because the template packages rules with remediation actions.
- If the question says new accounts must inherit checks, think organization rules or packs with recorder setup, because membership changes trigger deployment but prerequisites still apply.
- If the question says keep a central compliance dashboard and deploy checks, think an aggregatorA Config resource that collects configuration and compliance data from source accounts and Regions into a read-only view. plus an organization deployment, because aggregation is read-only.
- If the question says attach remediation directly to a pack rule, think update the pack template, because the pack owns its service-linked rules.
Remember
| Feature | Main job |
|---|---|
| Organization rule | Deploy an individual rule centrally |
| Organization conformance pack | Deploy rules and remediation together |
| Aggregator | Read configuration and compliance data centrally |
- Organization rule creation is API or CLI only.
- Organization rule and pack API deployments are per Region.
- Recorder unavailable in a newly joined account: deployment retries for seven hours.
- New accounts receive deployments; leaving accounts lose them. Account exclusions can limit pack deployment.
Trap: Adding remediation directly to an organization rule is unsupported by
PutRemediationConfigurations. Use a conformance pack template when deploying rules and remediation together.
Trap: Creating an aggregator does not deploy rules or establish recording in source accounts.
From the management or delegated administrator account, deploy the template to member accounts in the chosen Region.
The same organization deployment covers other included member accounts. Repeat deployment in each other required Region.
Box details (3)
- Config admin
Deploys the YAML pack through PutOrganizationConformancePack. A delegated administrator is a member account authorized to administer this integration.
- Member account A
Recording must be configured. The member cannot modify the centrally managed pack resources.
- Member account B
New accounts receive the pack after joining, subject to prerequisites and account exclusions.
From the AWS docs (17)
- A conformance pack is a collection of Config rules and remediation actions deployed as one entity to an account and Region or across an organization.
- A conformance pack uses a YAML template listing managed or custom Config rules and remediation actions.
- PutOrganizationConformancePack deploys conformance packs across member accounts and can be called only by the management account or a delegated administrator.
- Organization conformance pack deployment requires all features enabled in AWS Organizations.
- A delegated administrator for organization conformance packs must be registered for config-multiaccountsetup.amazonaws.com.
- Organization Config rules can be created only through the API or CLI, including PutOrganizationConfigRule, not the Config console.
- Organization rule and conformance pack API deployments are Region-specific and must be repeated in other target Regions.
- Organization rules and packs deploy to newly joined accounts and are removed when accounts leave the organization.
- AWS directs customers to enable Config recording before deploying organization rules and conformance packs.
- When a newly joined account lacks a recorder, deployment of existing organization rules and packs is retried for seven hours after the account joins.
- Member accounts cannot modify the underlying organization conformance pack rules and remediation actions.
- Organization conformance pack deployment can exclude specified accounts.
- To add remediation to a service-linked rule in a conformance pack, add it to the pack template and update the pack.
- PutRemediationConfigurations does not support service-linked rules, including organization rules, pack rules, and Security Hub rules.
- An aggregator replicates configuration and compliance data into a read-only view and cannot deploy rules to source accounts.
- Service-linked Config rules are owned by their creating AWS service and cannot be edited or deleted directly by a customer while subscribed to that service.
- A delegated administrator for an integrated AWS service is a member account registered to administer that service across the organization.
Aggregators and advanced queries
An AWS Config aggregatorA Config resource that collects configuration and compliance data from source accounts and Regions into a read-only view. copies configuration and complianceWhether a resource's configuration satisfies the conditions of a rule. data from source accounts and Regions into one read-only view. Advanced queryA property-based search of current recorded resource configurations using Config's supported SQL SELECT syntax. searches current recorded properties using a subset of SQLA language for storing, processing, and retrieving information in relational databases., a language for selecting and summarizing data. The aggregator supplies the data scopeThe resource types, identifiers, or tags used to decide which resource changes initiate a rule evaluation.; the query selects what to inspect.
How it works
- Enable Config recording in the source accounts and Regions. Choose the individual account sources or organization source for the aggregator.
- For individual sources, grant aggregation authorizationPermission for an aggregator account and Region to collect configuration and compliance data from an individual source account., permission for the aggregator account and Region to collect data. Organization sources do not require per-account authorization.
- The aggregator copies configuration and compliance data into the central account. It provides no write access back to sources and cannot deploy rules or enable their recording.
- In Advanced queries, select the aggregator as the query scope. Include
accountId,awsRegion, andresourceIdso each result identifies its source. For example, this query filters recorded EBS volume encryption properties:
SELECT accountId, awsRegion, resourceId, configuration.encrypted
WHERE resourceType = 'AWS::EC2::Volume'
AND configuration.encrypted = false
- Summarize current data with functions such as
COUNTor export console query results as comma-separated valuesA text format storing fields as comma-separated values. (CSV) or structured JSONA text data format representing structured information with named fields and values. data. For historical configurations, use the resource timelineThe sequence of recorded configuration items showing how a resource changed over time. or API configuration history. - For a QuickSight dashboard based on delivered files, AWS’s Config ResourceAn entity created and managed in AWS, such as an EC2 instance, storage volume, or security group. Compliance Dashboard uses a central S3 bucket receiving Config data. This is a separate reporting path from selecting an aggregator in Advanced queries.
Exam cues
- If the question says one view of compliance across accounts and Regions, think an aggregator, because it collects source configuration and compliance data.
- If the question says deploy checks to new organization accounts, think organization rules or conformance packsA template-defined bundle of Config rules and remediation actions deployed together., because an aggregator cannot deploy rules.
- If the question says find current recorded properties across accounts without separate service API calls, think an advanced query with aggregator scope, because it searches collected configuration itemsA point-in-time record of a supported resource's configuration, attributes, and relationships..
- If the question says individual source accounts have no aggregated data, think recording and authorization, because collection requires recorded source data and permission to read it.
- If the question says investigate yesterday’s resource settings, think configuration history, because advanced queries target current state.
Remember
| Need | Config path |
|---|---|
| Current properties in one account and Region | Local advanced query |
| Current properties across accounts and Regions | Advanced query with aggregator scope |
| Historical resource configurations | Resource timeline or configuration history API |
| Config data visualized in the documented QuickSight dashboard | Config file delivery to central S3 and the dashboard deployment |
- SQL support is a subset:
FROM,JOIN, andUNIONare unsupported. - Supported summary functions include
AVG,COUNT,MAX,MIN, andSUM. - Deleted resources and
ResourceNotRecordedCIs are not queryable. Recording support alone does not guarantee that every type or field is searchable. - An organization source does not need individual aggregation authorizations.
Trap: Data can be absent because it was never recorded. An aggregator does not create missing configuration history.
Trap: A central read-only compliance view cannot enforce or repair source configurations.
The aggregator collects recorded data from a source account and Region it is permitted to read.
It collects data from other included sources; organization aggregation does not require per-account authorization.
Advanced query reads the collected current configuration data. There is no deployment or write path back to the sources.
Box details (4)
- Source account A
Recording is enabled separately. Individual account sources authorize collection.
- Source account B
Organization sources do not need per-account aggregation authorization.
- Aggregator
Replicates configuration and compliance data without mutating the sources.
- Advanced query
Choose the aggregator as query scope and select accountId and awsRegion to identify each result's source.
From the AWS docs (17)
- A Config aggregator collects configuration and compliance data from multiple accounts and Regions, one account across Regions, or an organization whose accounts have Config enabled.
- Aggregators replicate data into a read-only view and cannot mutate source accounts, deploy rules, or push snapshots back to sources.
- Individually specified source accounts grant the aggregator account and Region permission to collect data; authorization is not required for organization sources.
- Aggregation collects data already recorded in Config; creating an aggregator does not enable recording in source accounts.
- Advanced queries use a subset of SQL SELECT to query current configuration items in one account and Region or through an aggregator across accounts and Regions.
- Advanced queries support AVG, COUNT, MAX, MIN, and SUM aggregation functions.
- Advanced queries do not support FROM, JOIN, or UNION and cannot query deleted resources or ResourceNotRecorded configuration items.
- Not every Config-supported resource type or configuration item field is searchable with advanced queries.
- In the query editor, selecting an aggregator as query scope runs the advanced query against that aggregator.
- Including accountId and awsRegion in an aggregator query identifies the source account and Region in the results.
- The Config console query editor can export results as CSV or JSON.
- Config's EBS volume query example selects resourceId, accountId, awsRegion, and configuration.encrypted for the AWS::EC2::Volume resource type.
- Resource configuration history is available through the console timeline and API.
- Organization conformance packs deploy rules and remediation actions across member accounts.
- The Config Resource Compliance Dashboard uses Config files delivered to a central S3 bucket and requires Amazon QuickSight for its visualizations.
- SQL is a language for storing, processing, and retrieving information in relational databases.
- Config advanced queries select properties with SELECT and can filter them with a WHERE condition.
Remediation controls and rule operations
AWS Config operational controls help manage automatic remediationA corrective action that Config runs through a Systems Manager Automation document for a resource reported as noncompliant. and interpret rule evaluations. Use them to limit remediation runs, investigate failures, and assess proposed resourceAn entity created and managed in AWS, such as an EC2 instance, storage volume, or security group. properties before deployment.
How it works
- For automatic remediation, supply
AutomationAssumeRole. A runbookA document defining the actions Systems Manager performs as a sequence of steps. parameter can hold a fixed value or receive the noncompliantWhether a resource's configuration satisfies the conditions of a rule. resource ID dynamically. For a service-linked ruleA Config rule owned and managed by the AWS service that created it. in a conformance packA template-defined bundle of Config rules and remediation actions deployed together., add remediation to the template and update the pack. See automatic remediation, roleAn assumable identity whose permissions determine which AWS actions it can perform. requirements, and service-linked rules. - Set the execution controls:
ConcurrentExecutionRatePercentagelimits the share of noncompliant resources remediated in parallel, whileErrorPercentagesets the allowed error percentage before Systems Manager stops automations for the rule. - Set
MaximumAutomaticAttemptsandRetryAttemptSecondsto control failed attempts within a counting window. When the threshold is reached within that window, Config adds a remediation exceptionAn exclusion preventing automatic remediation of a specified resource until the exception is cleared.. The exception blocks automatic remediation for that resource until cleared. See attempts, the retry window, and exceptions. - Investigate failed runs with
DescribeRemediationExecutionStatusordescribe-remediation-execution-status. Review execution states, step timestamps, failure messages, and parameter values. Check the Automation role’s action permissions and Systems Manager trust relationshipThe role policy specifying which identities or services may assume the role., plusiam:PassRolefor the caller starting Automation. See Automation troubleshooting. - For proactive evaluationEvaluation of proposed resource properties before deployment without blocking deployment or applying remediation., enable
PROACTIVEon a supporting rule, callStartResourceEvaluationwith resource properties, then pass the returnedResourceEvaluationIdtoGetResourceEvaluationSummary. Only supporting managed rulesA predefined Config compliance check created by AWS with customizable settings. and resource types can be used. The pipeline must enforce any rejection because proactive rules do not block deployment or remediate failures.
Exam cues
- If the question says a compliant resource was remediated, think a stale compliance snapshotThe periodically captured compliance results that Config uses to select resources for automatic remediation., because automatic remediation can use a snapshot captured before the resource became compliant.
- If the question says remediate public S3 bucket access, think
AWS-DisableS3BucketPublicReadWrite, because it uses S3 Block Public Access and takes the bucket inS3BucketName. SupplyAutomationAssumeRolefor automatic remediation. - If the question says choose a rotation age in
cmk-backing-key-rotation-enabled, think no parameter, because this periodicA rule trigger that starts evaluation at a selected frequency. rule checks whether automatic KMS key rotation is enabled and has no inputs. - If the question says remediate
required-tagswithAWS-SetRequiredTags, think a custom Systems Manager Automation document, because the managed document does not work with this rule. The rule checks up to six tagsA label with a key and an optional value used to describe a resource. and does not prevent resource creation. - If the question says require EC2 Dedicated HostsA physical server with EC2 instance capacity dedicated to your use., think
Tenancy=host, becausededicatedis a distinct value.ModifyInstancePlacementacceptsdefault,dedicated, andhost. - If the question asks for evaluation results, think rule status:
COMPLIANTmeans pass,NON_COMPLIANTmeans fail,ERRORmeans invalid parameters, andNOT_APPLICABLEmeans the rule logic does not apply.
Remember
ConcurrentExecutionRatePercentagedefaults to 10 percent;ErrorPercentagedefaults to 50 percent.MaximumAutomaticAttemptsdefaults to 5 and accepts 1 through 25.RetryAttemptSecondsdefaults to 60 seconds.- A remediation exception blocks automatic remediation until cleared.
Trap: A proactive
NON_COMPLIANTresult does not stop deployment or fix the resource. A pipeline must enforce its own rejection decision.
From the AWS docs (21)
- Automatic remediation uses periodically captured compliance snapshots and can run on resources that became compliant after the snapshot.
- DescribeRemediationExecutionStatus and its CLI equivalent describe-remediation-execution-status report execution states, step timestamps, and failure messages.
- A remediation parameter can use a static value or dynamically receive the noncompliant resource ID.
- Automation failures can result from insufficient action permissions, a missing Systems Manager trust relationship, or missing iam:PassRole permission for the caller starting Automation.
- ConcurrentExecutionRatePercentage limits the percentage of noncompliant resources remediated in parallel and defaults to 10 percent.
- ErrorPercentage controls the error percentage allowed before Systems Manager stops running automations for the rule and defaults to 50 percent.
- MaximumAutomaticAttempts is the maximum failed automatic-remediation attempts, defaults to 5, and accepts 1 through 25.
- RetryAttemptSeconds is the time window for counting attempts before adding a remediation exception and defaults to 60 seconds.
- AWS-DisableS3BucketPublicReadWrite uses S3 Block Public Access and takes the target bucket in S3BucketName.
- Automatic Config remediation requires a value for AutomationAssumeRole.
- Add remediation for a service-linked pack rule to the conformance pack template and update the pack.
- A remediation exception excludes a resource from automatic remediation until the exception is cleared.
- After enabling PROACTIVE evaluation on a supporting rule, call StartResourceEvaluation with resource properties and use its ResourceEvaluationId with GetResourceEvaluationSummary to retrieve the result.
- Only supporting resource types and managed rules can be used for proactive evaluation.
- Proactive rules neither prevent deployment nor remediate resources flagged as NON_COMPLIANT.
- The periodic cmk-backing-key-rotation-enabled rule checks whether automatic KMS key rotation is enabled and has no input parameters.
- The configuration-change required-tags rule checks specified resource tags, supports up to six tags, and does not prevent resource creation.
- AWS-SetRequiredTags does not work as remediation for required-tags; that rule requires a custom Systems Manager Automation document for remediation.
- Rule results include COMPLIANT for passing, NON_COMPLIANT for failing, ERROR for invalid parameters, and NOT_APPLICABLE when rule logic does not apply.
- Launching an instance on a Dedicated Host uses Tenancy=host; dedicated is a distinct tenancy value.
- ModifyInstancePlacement accepts default, dedicated, and host for its Tenancy parameter.
Flashcards
Nothing to review here right now. Pick other topics or another filter.
·
All 51 cards as a list
What does AWS Config configuration history contain, and how far back does it reach?
Configuration items describing supported resource configurations and relationships, from when recording started. Use the resource timeline or API to review them.
An AWS Config configuration-change rule has no evaluations for an EC2 resource. What recording settings should an operator inspect?
Check that the recorder is running, includes the resource type, and produces changes matching the rule's scope. Periodic rules are different and can run without configuration items.
How do AWS Config Continuous and Daily recording differ, and which does Firewall Manager need?
Continuous captures changes as they occur. Daily captures the latest state over 24 hours if different from the previous item. Firewall Manager depends on continuous recording.
How should an AWS Config deployment avoid duplicate records of global IAM users, groups, roles, and customer managed policies?
Record those types in one Region where Config was supported before February 2022. Regions added after that date cannot record those global IAM types.
An auditor needs historical Systems Manager patch and State Manager association compliance. What should be recorded in AWS Config?
Record AWS::SSM::PatchCompliance and AWS::SSM::AssociationCompliance. Systems Manager shows current compliance; Config provides its history and change tracking.
An auditor needs configuration evidence for AWS Organizations accounts, organizational units, and policies. Can AWS Config record those resource types?
Yes. Supported types include Account, Organization, OrganizationalUnit, Policy, and ResourcePolicy under AWS::Organizations.
How do AWS Config configuration history and configuration snapshots differ, and where are their files delivered?
History follows changes to a resource over time; a snapshot collects recorded resource configurations together. Config delivers files to the configured S3 bucket, with history files every six hours when changes occur.
Can an AWS Config periodic rule evaluate when its configuration recorder is disabled?
Yes. Periodic rules do not depend on configuration items. Do not apply the recording prerequisite for configuration-change rules to every rule.
An organization wants to protect baseline AWS Config recorders and delivery channels while retaining other permitted Config administration. Which permissions should an SCP deny?
Use an SCP denying config:DeleteConfigurationRecorder, config:StopConfigurationRecorder, config:PutConfigurationRecorder, config:DeleteDeliveryChannel, and config:PutDeliveryChannel. Target these actions rather than config:*; AWS documents the pattern in its Control Tower Config protection SCP.
When does an AWS Config rule evaluate, and how can its configuration-change scope be restricted?
Config evaluates when the rule is added, then on configuration changes, a periodic schedule, or both. Change-trigger scope can select resource types, a type and ID, a tag key and value, or all recorded changes.
What are the implementation choices for AWS Config managed and custom rules?
Managed rules are AWS-defined checks with configurable parameters. Custom rules use Lambda code that reports results with PutEvaluations, or a Guard policy without a custom Lambda function.
A team needs an AWS Config check expressed as comparisons of recorded properties without developing a Lambda function. Which custom rule type fits?
A custom policy rule written in Guard. Detective Guard evaluations are triggered by configuration changes.
How do detective and proactive AWS Config evaluations differ?
Detective checks deployed resources. Proactive checks proposed properties before deployment, but does not block deployment or remediate failures.
A deployment pipeline uses AWS Config proactive evaluation and receives NON_COMPLIANT. What must stop deployment?
A decision enforced by the pipeline. Config returns an evaluation result; proactive rules do not prevent deployment themselves.
A team needs AWS Config to find running EC2 instances launched from images outside an approved AMI ID list. Which rule and parameter fit?
Use approved-amis-by-id with amiIds. It evaluates configuration changes and marks running instances with unlisted AMIs NON_COMPLIANT.
A team needs AWS Config to find attached EBS volumes without encryption or using an unapproved KMS key. What should it use?
Use encrypted-volumes and optionally set kmsId to the required key. Its scope is attached EBS volumes, and its trigger is configuration changes.
A team wants AWS Config to detect when no CloudTrail trail is enabled. Which managed rule fits, and does it run on every change?
Use cloudtrail-enabled. It is periodic, so its evaluation is scheduled rather than triggered by every configuration change.
A security team wants AWS Config managed and custom rule evaluations alongside Security Hub CSPM findings. Which integration is needed?
Activate the AWS Config integration. It sends evaluations through EventBridge; only evaluations performed after Security Hub CSPM was enabled are sent. Service-linked rules are excluded except Control Tower rules.
Which AWS Config managed rules check S3 buckets for public read access, public write access, or public bucket-level settings with an exclusion list?
Use s3-bucket-public-read-prohibited, s3-bucket-public-write-prohibited, and s3-bucket-level-public-access-prohibited, respectively. The first two check Block Public Access, bucket policy, and bucket ACL; the last accepts excludedPublicBuckets.
A team wants AWS Config to detect old active IAM user access keys, alert operators, and support an automated response. Which components fit?
Use access-keys-rotated with maxAccessKeyAge, an EventBridge compliance-change rule targeting SNS, and a suitable Automation remediation runbook if a fix is required. The age defaults to 90 days; root keys are excluded.
How does an AWS Config rule automatically correct a noncompliant resource?
Configure automatic remediation with a Systems Manager Automation runbook, its parameters, and AutomationAssumeRole. A parameter can receive the resource ID dynamically.
What starts manual versus automatic AWS Config remediation?
An operator starts manual remediation for selected resources. Automatic remediation starts for reported noncompliance and can retry failed actions within its configured window.
A security team needs AWS Config rules and automatic fixes for unwanted S3 public access across member accounts. How should it package the deployment?
Use an organization conformance pack containing the rule and suitable Automation remediation. Add pack-rule remediation through the pack template, rather than attaching it directly to an organization rule.
A team wants an alert when AWS Config detects unrestricted incoming SSH in security groups, but wants a person to decide on the fix. What should it configure?
Use restricted-ssh and an EventBridge compliance-change rule targeting SNS. Keep remediation manual so an operator starts any configured fix.
An AWS Config compliance alert needs to notify an SNS topic without automatically correcting resources. Which event path fits?
Match aws.config Config Rules Compliance Change events with EventBridge and target SNS. Automatic remediation is a separately configured action; Config event delivery is best effort.
A company must turn on S3 server access logging automatically for any bucket that lacks it. What fits?
The AWS Config rule s3-bucket-logging-enabled with automatic remediation through the AWS-ConfigureS3BucketLogging runbook.
An AWS Config audit needs every configuration change for selected supported resources. Should the team choose daily or continuous recording?
Continuous recording. Daily recording retains only the latest changed state over a 24-hour period.
What does an AWS Config conformance pack contain, and what format defines it?
A YAML template bundling managed or custom Config rules and remediation actions. Deploy it as one entity in an account and Region or across an organization.
How does an AWS Config organization conformance pack follow account membership changes?
It deploys to newly joined accounts and is removed when accounts leave. Deploy from the management account or a delegated administrator; member accounts cannot modify its rules and remediation actions.
How are AWS Config organization rules created, and what happens when a newly joined account has no recorder?
Create rules through the API or CLI, such as PutOrganizationConfigRule, separately in each Region. Deployment retries for seven hours after the account joins if a recorder is unavailable.
A security team needs common AWS Config checks deployed across member accounts and one central compliance view. Which two features fit?
An organization conformance pack deploys the rules and remediation actions. An aggregator collects a read-only central view; it does not deploy the checks.
A new AWS account joins an organization with AWS Config organization rules, but no recorder exists. What should the administrator do within the deployment retry window?
Create and enable the required recorder. Existing organization rule and pack deployments are retried for seven hours after the account joins when a recorder is unavailable.
A team needs to add automatic remediation to an AWS Config rule created by a conformance pack. Should it call PutRemediationConfigurations directly?
No. Add the remediation action to the conformance pack template and update the pack. Direct remediation configuration is unsupported for these service-linked rules.
What does an AWS Config aggregator collect, and when is source-account authorization needed?
It collects configuration and compliance data across accounts and Regions. Individual source accounts authorize the aggregator account and Region; organization sources do not need that authorization.
Does creating an AWS Config aggregator enable source recorders or deploy compliance rules?
No. Enable recording separately and use organization deployments for rules or packs. Aggregation is read-only.
A security team needs a list of recorded unencrypted EBS volumes across accounts and Regions. Which AWS Config feature fits?
Run an advanced query with the aggregator as scope. Select source account and Region with the volume ID, and filter the recorded encryption property.
A team creates an AWS Config aggregator for individually listed source accounts but sees no data. What setup should it inspect?
Check source recording and permission granted to the aggregator account and Region. Individual sources need aggregation authorization; organization sources do not.
An auditor uses an AWS Config advanced query to find deleted resources or yesterday's configurations. Is this the right query path?
No. Advanced queries target current recorded state and exclude deleted resources. Use configuration history through the resource timeline or API for historical configurations.
A team wants an Amazon QuickSight dashboard built from AWS Config compliance evidence. What data delivery does AWS's Config Resource Compliance Dashboard require?
Enable Config in the accounts and Regions being tracked and deliver its files to a centralized S3 bucket. The dashboard uses that data and QuickSight for visualization.
AWS Config automatic remediation starts for a resource that is already compliant. Why can this happen?
Its compliance snapshot can be stale. A resource corrected after the snapshot can still be selected for automatic remediation.
AWS Config remediation fails despite a valid runbook. Which execution details and permissions should an operator inspect?
Use DescribeRemediationExecutionStatus for failed steps and errors. Check parameter values, Automation role action permissions and Systems Manager trust, and any PassRole error for the caller starting Automation.
Which AWS Config remediation controls limit simultaneous fixes and repeated failures?
ConcurrentExecutionRatePercentage limits parallel remediation; ErrorPercentage controls when errors stop further automations. MaximumAutomaticAttempts and RetryAttemptSeconds control retries before an exception.
An AWS Config rule flags unwanted public S3 bucket access. Which runbook and parameters can support automatic correction?
AWS-DisableS3BucketPublicReadWrite uses S3 Block Public Access. Supply S3BucketName and AutomationAssumeRole for automatic Config remediation; pack rules require remediation in the pack template.
What are AWS Config automatic remediation's default failed-attempt count and retry window?
MaximumAutomaticAttempts defaults to 5 and RetryAttemptSeconds to 60 seconds. Reaching the failure threshold within the window adds a remediation exception that blocks further automatic remediation.
What are the AWS Config remediation defaults for concurrency and allowed errors?
ConcurrentExecutionRatePercentage defaults to 10 percent; ErrorPercentage defaults to 50 percent.
An AWS Config resource has a remediation exception after repeated failures. Will automatic remediation keep retrying it?
No. The exception excludes it from automatic remediation until the exception is cleared.
Which AWS Config API sequence lets a pipeline assess proposed resource properties before deployment?
Enable proactive mode on a supporting rule, call StartResourceEvaluation, then retrieve GetResourceEvaluationSummary using the returned ResourceEvaluationId. The pipeline must enforce any rejection.
Can the AWS Config cmk-backing-key-rotation-enabled rule enforce a chosen rotation age through a parameter?
No. It checks whether automatic rotation is enabled, runs periodically, and has no input parameters.
A team uses AWS Config required-tags and wants to attach AWS-SetRequiredTags as remediation. Is that supported?
No. Use a custom Systems Manager Automation document for this rule. required-tags checks up to six tags and does not prevent creation of incorrectly tagged resources.
What do COMPLIANT, NON_COMPLIANT, ERROR, and NOT_APPLICABLE mean for an AWS Config rule?
Pass, fail, invalid rule parameters, and rule logic that does not apply to the resource, respectively.
A custom AWS Config rule must require EC2 placement on Dedicated Hosts. Is placement tenancy dedicated sufficient?
No. Check placement tenancy for host. dedicated and host are different EC2 tenancy values; launching on a Dedicated Host uses Tenancy=host.