Cloud concept · Updated

Data Classification - Data classification is the process of categorizing information by sensitivity, business impact, and handling requirements so that appropriate security controls can be applied.

In short

Classification connects the value and risk of information to decisions about access, storage, monitoring, retention, and disposal. In cloud workloads, it helps engineers determine which activities require audit records, which changes should trigger alerts, and how quickly security teams must respond.

Cloud Concepts Hub →

6 key aspects of Data Classification

Effective classification combines accountable ownership, workload discovery, explicit handling rules, and risk-proportionate monitoring throughout the data lifecycle.

01
Defined categories and ownership

Use documented classification criteria and assign a data owner who can approve the category and handling requirements. Consistent definitions prevent teams from treating the same type of information differently.

02
Workload and data-flow assessment

Identify sensitive information across primary stores, processing paths, exports, backups, and logs. Include confidentiality, integrity, and availability impacts when determining monitoring requirements.

03
Classification-driven monitoring

Map each category to required audit events, detection latency, retention, and response ownership. Verify that telemetry captures relevant data access as well as administrative changes.

04
Discovery with human validation

Use automated inspection to locate potentially sensitive content, then validate results against business context. Track inspection gaps and do not interpret an absence of findings as proof that data is nonsensitive.

05
Labels linked to enforceable controls

Maintain classification metadata in an authoritative inventory and connect it to access policies and monitoring configurations. Tags and catalog labels provide context but enforce nothing unless an applicable control uses them.

06
Lifecycle review and proportionality

Reassess classifications when data is copied, combined, transformed, or used for a new purpose. Balance monitoring depth against cost and noise while preserving required evidence and protecting the logs themselves.

Cloud services for Data Classification

Macie helps identify potentially sensitive S3 content, while data owners validate its business classification. Lake Formation can translate approved classifications into governed data lake access policies. CloudTrail supplies audit evidence for supported administrative and data-access activity, and CloudWatch supports workload telemetry and alarms. AWS Config evaluates resource configurations against handling requirements. EventBridge routes supported findings and events into response workflows. Together, these services support discovery, enforcement, detection, and review, but engineers must explicitly configure coverage and connect classification metadata to monitoring decisions.

Amazon MacieAmazon Macie

Macie discovers potentially sensitive data in supported Amazon S3 objects using managed and custom data identifiers. Its findings inform classification reviews and monitoring priorities, but do not replace business-owner decisions or cover every cloud data store.

AWS CloudTrailAWS CloudTrail

CloudTrail records supported AWS API activity for auditing and investigation. Configure relevant data events explicitly when classified workloads require visibility into data-plane activity such as S3 object access; management events alone are insufficient.

Amazon CloudWatchAmazon CloudWatch

CloudWatch collects workload logs and metrics and supports alarms for classification-driven monitoring requirements. For example, log metric filters can turn selected audit events into metrics that trigger alarms, provided the required events are delivered to CloudWatch Logs.

AWS ConfigAWS Config

AWS Config records supported resource configurations and evaluates them against configured rules. Engineers can use it to assess whether resources holding classified data meet requirements such as restricted exposure or required tags, but it does not inspect data content or audit individual reads.

Amazon EventBridgeAmazon EventBridge

EventBridge routes supported security findings and operational events to notification or response targets. Classification context can inform routing and severity through event content or an enrichment workflow; EventBridge does not classify datasets itself.

AWS Lake FormationAWS Lake Formation

Lake Formation supports governed access to data lake resources, including tag-based access control using LF-Tags. A classification taxonomy can inform those tags and permissions, but policies and supported access paths must be configured for enforcement.

Data Classification exam questions

Real practice exam questions on Data Classification, with the correct answer and its explanation. Answer them with every option in the exam simulator.

SOA-C03

A company stores reports smaller than 1 GB in an Amazon S3 general purpose bucket. A trusted governance process assigns each report a Classification object tag of Internal or Restricted. Each report also has a Project object tag.

An archive application uses CopyObject to copy reports to another S3 bucket. The application sets the destination object's Project tag to Archive by using the REPLACE tagging directive, but supplies no other tags. An audit finds that archived reports have lost their classification labels.

The application must preserve each report's approved classification and set its Project object tag to Archive. Source objects must remain unchanged, and the destination labels must be available as S3 object tags for tag-based access policies. No process modifies source tags during copying.

Which change should a CloudOps engineer make?

Show answerCORRECT ANSWER

Retrieve the source object's tags, retain its Classification value, set Project to Archive in the retrieved tag set, and supply the resulting tag set with the REPLACE tagging directive.

The REPLACE tagging directive replaces the complete destination object tag set rather than merging supplied tags with source tags. Retrieving and updating the source tag set preserves the approved Classification while assigning Project=Archive, without modifying the source object.

SCS-C03

A healthcare company is designing monitoring for a claims-processing application. Diagnostic events contain patient names, diagnoses, request IDs, processing outcomes, and latency measurements. Investigators require the original diagnostic events, but operations personnel need only request IDs, outcomes, and latency measurements.

Company policy requires patient-identifiable information to remain in a dedicated security account. An approved schema containing only request IDs, outcomes, and latency measurements is classified as Internal and can be stored in a shared operations account. Request IDs cannot be resolved to patients by operations personnel.

The proposed design sends complete diagnostic events to Amazon CloudWatch Logs and uses subscription filters to deliver telemetry to the operations account. The security engineer must establish appropriate monitoring requirements for the events and their downstream copies.

Which design meets the requirements?

Show answerCORRECT ANSWER

Classify the complete diagnostic stream as Restricted and store it only in the security account. Have the application emit a separate stream that contains only fields permitted by the approved Internal schema, and deliver that stream to the operations account.

This design applies handling requirements to the actual contents of each stream. Original patient-identifiable events remain available to investigators within the required security boundary, while operations receives only an explicitly approved, lower-sensitivity projection. Copies do not acquire a lower classification merely because they are used for monitoring.

DEA-C01

A financial services company stores customer records in an Amazon RDS for PostgreSQL database. Before approving a new analytics export, a data steward needs a report identifying columns that contain supported types of personally identifiable information (PII), including email addresses and credit card numbers. Column names are not reliable indicators of their contents. Security policy permits a processing job to inspect database records in memory, but prohibits writing record values to Amazon S3. Only column names and detected PII categories may be retained in the report. Which solution meets these requirements?

Show answerCORRECT ANSWER

Run an AWS Glue ETL job that reads the database through JDBC and uses sensitive data detection to identify columns containing the required PII types. Write only column names and detected PII categories to the report, without writing or logging record values.

AWS Glue can read JDBC sources and perform content-based sensitive data detection, including column identification for supported PII types. The job can retain only classification metadata, meeting the reporting requirement without staging database records in S3.

Practice Data Classification in a full exam

Questions on Data Classification in each practice exam, with every answer option and an explanation for each:

Start a free practice exam

What is Data Classification

3 min read

Data classification assigns information to defined categories according to the consequences of unauthorized disclosure, modification, loss, or unavailability. Organizations commonly use labels such as public, internal, confidential, and restricted, but these names and their meanings are organization-specific. A useful classification policy defines the criteria for each category, identifies accountable data owners, and maps categories to concrete handling and monitoring requirements.

Start by understanding the workload rather than selecting logging services first. Identify what information it processes, where that information enters and leaves, who can access it, and which legal, contractual, or business obligations apply. Examine databases, object stores, message queues, backups, analytics exports, and application logs. A customer profile database and a public documentation website need different confidentiality monitoring, while the public website may still need strong integrity monitoring to detect unauthorized changes.

Apply classifications at a level that supports meaningful controls. A database table might contain restricted identifiers alongside ordinary operational attributes. Where practical, separate sensitive fields or datasets so access policies and monitoring can be targeted. Record classifications in a governed catalog, resource tags, object metadata, or another authoritative inventory. Resource-level labels are useful context, but a bucket tag does not prove that every object has the same sensitivity. Copies, temporary exports, and derived datasets also need assessment.

Translate classifications into explicit monitoring requirements. For a restricted payroll dataset, engineers might require records of object reads and writes, permission changes, public-access changes, deletion attempts, and changes to encryption-key policies. Correlate these events with the requesting identity, source network, workload, and expected access patterns. Escalate an unexpected external grant more urgently than a routine authorized read. Define alert owners, response time targets, log retention, and acceptable detection latency before implementing alarms.

In AWS, distinguish control-plane activity from access to the data itself. CloudTrail management events can reveal many configuration and permission changes, but they do not provide comprehensive records of object reads. Relevant CloudTrail data events, such as Amazon S3 object-level activity, must be explicitly configured. Database query auditing and application-level access logging may require separate mechanisms. Organization-wide collection should cover the accounts and Regions containing classified information, with centralized protection against unauthorized log alteration or deletion.

Automated discovery can identify patterns such as personal identifiers and help locate unexpectedly sensitive content. It cannot reliably establish every business classification: an unreleased acquisition plan may be highly confidential without matching a standard identifier pattern. Validate discovery findings with data owners, review unsupported or uninspected content, and reassess classifications when schemas, integrations, or business uses change. Encryption does not reduce the underlying classification, and a classification label does not itself enforce authorization.

The main trade-off is balancing visibility against cost, operational noise, and additional exposure. Detailed access logging can generate substantial event volume, while application logs can accidentally capture credentials or personal information. Prefer useful audit metadata over unnecessary payloads, restrict access to monitoring data, and choose retention according to investigative and compliance needs. Maintain baseline monitoring for every workload, then increase depth and urgency based on classification, exposure, and business criticality rather than assuming that all data requires identical treatment.

Learn Data Classification in lessons

5 short lessons on Data Classification in 5 certification courses.

Open the lessons