Cette page n'est pas encore disponible en français, sa traduction est en cours.
Si vous avez des questions ou des retours sur notre projet de traduction actuel,
n'hésitez pas à nous contacter.
Overview
Sensitive data, such as credit card numbers, bank routing numbers, and API keys are often exposed unintentionally in application logs, APM spans, and RUM events, which can expose your organization to financial and privacy risks.
Sensitive Data Scanner is a stream-based, pattern matching service used to identify, tag, and optionally redact or hash sensitive data. Security and compliance teams can implement Sensitive Data Scanner as a new line of defense, helping prevent against sensitive data leaks and limiting non-compliance risks.
To use Sensitive Data Scanner, set up a scanning group to define what data to scan and then set up scanning rules to determine what sensitive information to match within the data.
This document walks you through the following:
- The permissions required to view and set up Sensitive Data Scanner.
- Setting up scanning for sensitive data.
- Using the out-of-the-box dashboard.
Note: See PCI DSS Compliance for information on setting up a PCI-compliant Datadog organization.
Set up Sensitive Data Scanner
There are two locations where you can redact your sensitive data:
In the cloud:
- With Sensitive Data Scanner in the Cloud, you submit your logs in the Datadog backend. In this method, logs leave your premises before they are redacted. You can have multiple scanning groups per organization, and you can create custom scanning rules. You can also redact sensitive data in tags.
In your environment:
Join the Preview!
Sensitive Data Scanner support for the Datadog Agent is in Preview. To enroll, click Request Access.
Request Access- With Sensitive Data Scanner using the Agent, Datadog redacts your logs before submitting them to the Datadog backend, and unredacted logs never need to leave your premises. With this method, you are limited to one scanning group per organization, and you can use only predefined library rules.
With Observability Pipelines:
When you set up a pipeline in Observability Pipelines, add the Sensitive Data Scanner processor to redact sensitive data in your logs before they leave your environment. See Observability Pipelines for more information.
Prerequisites
By default, users with the Datadog Admin role have access to view and set up scanning rules. To allow other users access, grant the data_scanner_read
or data_scanner_write
permissions under Compliance to a custom role. See Access Control for details on how to set up roles and permissions.
Grant appropriate permissions. By default, users with the Datadog Admin role have access to view and set up scanning rules. To allow other users access, grant the data_scanner_read
or data_scanner_write
permissions under Compliance to a custom role. See Access Control for details on how to set up roles and permissions.
Follow the steps to enable remote configuration.
Install the Datadog Agent v7.54 or newer.
Add a scanning group
A scanning group determines what data to scan. It consists of a query filter and a set of toggles to enable scanning for logs, APM, RUM, and events. See the Log Search Syntax documentation to learn more about query filters.
For Terraform, see the Datadog Sensitive Data Scanner group resource.
To set up a scanning group, perform the following steps:
- Navigate to the Sensitive Data Scanner configuration page.
- Click Add scanning group. Alternatively, click the Add dropdown menu on the top right corner of the page and select Add Scanning Group.
- Enter a query filter for the data you want to scan. At the top, click APM Spans to preview the filtered spans. Click Logs to see the filtered logs.
- Enter a name and description for the group.
- Click the toggle buttons to enable Sensitive Data Scanner for the products you want (for example, logs, APM spans, RUM events, and Datadog events).
- Click Create.
Note: Sensitive Data Scanner using the Agent supports only one scanning group per organization.
A scanning group determines what logs to scan. It consists of a query filter to match eligible agents based on host tags.
To set up a scanning group, perform the following steps:
- Navigate to the Sensitive Data Scanner using the Agent configuration page.
- Click Add scanning group. Alternatively, click the Add dropdown menu on the top right corner of the page and select Add Scanning Group.
- Enter a query filter for the data you want to scan. You can use only host-level tags for matching agents. At the bottom, the number of matching and eligible agents is displayed, including the total number out of all agents that match the tag.
- Enter a name and description for the group.
- Click Save.
By default, a newly-created scanning group is disabled. To enable a scanning group, click the corresponding toggle on the right side.
Add scanning rules
A scanning rule determines what sensitive information to match within the data defined by a scanning group. You can add predefined scanning rules from Datadog’s Scanning Rule Library or create your own rules using regex patterns. The data is scanned at ingestion time during processing. For logs, this means the scan is done before indexing and other routing decisions.
For Terraform, see the Datadog Sensitive Data Scanner rule resource.
To add scanning rules, perform the following steps:
- Navigate to the Sensitive Data Scanner configuration page.
- Click the scanning group where you want to add the scanning rules.
- Click Add Scanning Rule. Alternatively, click the Add dropdown menu on the top right corner of the page and select Add Scanning Rule.
- Select whether you want to add a library rule or create a custom scanning rule.
The Scanning Rule Library contains predefined rules for detecting common patterns such as email addresses, credit card numbers, API keys, authorization tokens, and more.
- Select a scanning group if you did not create this rule within a scanning group.
- In the Add library rules to the scanning group section, select the library rules you want to use.
- In the Define rule target and action section, select if you want to scan the Entire Event or Specific Attributes.
- If you are scanning the entire event, you can optionally exclude specific attributes from getting scanned.
- If you are scanning specific attributes, specify which attributes you want to scan.
- For Define actions on match, select the action you want to take for the matched information. Note: Redaction, partial redaction, and hashing are all irreversible actions.
- Redact: Replaces all matching values with the text you specify in the Replacement text field.
- Partially Redact: Replaces a specified portion of all matched data. In the Redact section, specify the number of characters you want to redact and which part of the matched data to redact.
- Hash: Replaces all matched data with a unique identifier. The UTF-8 bytes of the match is hashed with the 64-bit fingerprint of FarmHash.
- Optionally, add tags you want to associate with events where the values match the specified regex pattern. Datadog recommends using
sensitive_data
and sensitive_data_category
tags. These tags can then be used in searches, dashboards, and monitors. See Control access to logs with sensitive data for information on how to use tags to determine who can access logs containing sensitive information. - For Set priority level, select the priority level for the rule based on your business needs.
- In the Name and describe the scanning rule section, enter a name for the rule. Optionally, add a description.
- Click Add Rules.
Add additional keywords
After adding OOTB scanning rules, you can edit each rule separately and add additional keywords to the keyword dictionary.
- Navigate to the Sensitive Data Scanner configuration page.
- Click the scanning group with the rule you want to edit.
- Hover over the rule, and then click the pencil icon.
- The recommend keywords are used by default. To add additional keywords, toggle Use recommended keywords, then add your keywords to the list. You can also require that these keywords be within a specified number of characters of a match. By default, keywords must be within 30 characters before a matched value.
- Click Update.
You can create custom scanning rules using regex patterns to scan for sensitive data.
- Select a scanning group if you did not create this rule within a scanning group.
- In the Define match conditions section, specify the regex pattern to use for matching against events in the Define the regex field. Enter sample data in the Add sample data field to verify that your regex pattern is valid.
Sensitive Data Scanner supports Perl Compatible Regular Expressions (PCRE), but the following patterns are not supported:
- Backreferences and capturing sub-expressions (lookarounds)
- Arbitrary zero-width assertions
- Subroutine references and recursive patterns
- Conditional patterns
- Backtracking control verbs
- The
\C
“single-byte” directive (which breaks UTF-8 sequences) - The
\R
newline match - The
\K
start of match reset directive - Callouts and embedded code
- Atomic grouping and possessive quantifiers
- For Create keyword dictionary, add keywords to refine detection accuracy when matching regex conditions. For example, if you are scanning for a sixteen-digit Visa credit card number, you can add keywords like
visa
, credit
, and card
. You can also require that these keywords be within a specified number of characters of a match. By default, keywords must be within 30 characters before a matched value. - In the Define rule target and action section, select if you want to scan the Entire Event or Specific Attributes.
- If you are scanning the entire event, you can optionally exclude specific attributes from getting scanned.
- If you are scanning specific attributes, specify which attributes you want to scan.
- For Define actions on match, select the action you want to take for the matched information. Note: Redaction, partial redaction, and hashing are all irreversible actions.
- Redact: Replaces all matching values with the text you specify in the Replacement text field.
- Partially Redact: Replaces a specified portion of all matched data. In the Redact section, specify the number of characters you want to redact and which part of the matched data to redact.
- Hash: Replaces all matched data with a unique identifier. The UTF-8 bytes of the match is hashed with the 64-bit fingerprint of FarmHash.
- Optionally, add tags you want to associate with events where the values match the specified regex pattern. Datadog recommends using
sensitive_data
and sensitive_data_category
tags. These tags can then be used in searches, dashboards, and monitors. See Control access to logs with sensitive data for information on how to use tags to determine who can access logs containing sensitive information. - For Set priority level, select the priority level for the rule based on your business needs.
- In the Name and describe the scanning rule section, enter a name for the rule. Optionally, add a description.
- Click Add Rule.
Notes:
- Any rules that you add or update affect only data coming into Datadog after the rule was defined.
- Sensitive Data Scanner does not affect any rules you define on the Datadog Agent directly.
- After rules are added, ensure that the toggles for your scanning groups are enabled to begin scanning.
See Investigate Sensitive Data Issues for details on how to use the Summary page to triage your sensitive data issues.
A scanning rule determines what sensitive information to match within the data defined by a scanning group. The Datadog Agent scans your data in your local environment during log collection, before logs are sent to the Datadog platform.
Note: Sensitive Data Scanner using the Agent supports only predefined scanning rules from Datadog's Scanning Rule Library. The total number of scanning rules is limited to 20.
To add scanning rules, perform the following steps:
- Navigate to the Sensitive Data Scanner using the Agent configuration page.
- Click Add Scanning Rule. Alternatively, click the Add dropdown menu on the top right corner of the page and select Add Scanning Rule.
- In the Add library rules to the scanning group section, select the library rules you want to use. Use the Filter library rules input to search existing library rules. Next to the rule name you can find the list of predefined tags for each rule.
- In the Define rule target and action section, select the action that you want to take for the matched sensitive information. Note: Redaction, partial redaction, and hashing are all irreversible actions.
- Redact: Replaces all matching values with the text you specify in the Replacement text field.
- Partially Redact: Replaces a specified portion of all matched data. In the Redact section, specify the number of characters you want to redact and which part of the matched data to redact.
- Hash: Replaces all matched data with a unique identifier. The UTF-8 bytes of the match is hashed with the 64-bit fingerprint of FarmHash.
- Optionally, add additional tags you want to associate with events where the values match the specified regex pattern. Datadog adds
sensitive_data
and sensitive_data_category
tags by default. These tags can then be used in searches, dashboards, and monitors. See Control access to logs with sensitive data for information on how to use tags to determine who can access logs containing sensitive information. - Click Save.
Notes:
- Any rules that you add or update affect only data coming into Datadog after the rule was defined.
- After rules are added, ensure that the toggles for your scanning groups are enabled to begin scanning.
Excluded namespaces
There are reserved keywords that the Datadog platform requires for functionality. If any of these words are in a log that is being scanned, the 30 characters after the matched word are ignored and not redacted. For example, what comes after the word date
in a log is usually the event timestamp. If the timestamp is accidentally redacted, that would result in issues with processing the log and being able to query it later. Therefore, the behavior for excluded namespaces is to prevent unintentionally redacting important information for product functionality.
The excluded namespaces are:
host
hostname
syslog.hostname
service
status
env
dd.trace_id
trace_id
trace id
dd.span_id
span_id
span id
@timestamp
timestamp
_timestamp
Timestamp
date
published_date
syslog.timestamp
error.fingerprint
x-datadog-parent-id
Edit scanning rules
Navigate to the Sensitive Data Scanner configuration page.
Hover over the scanning rule you want to edit and click the Edit (pencil) icon.
The Define match conditions section shows either the regular expression you wrote for your custom rule or an explanation of the library scanning rule you chose along with examples of matched sensitive information.
To make sure that a rule matches your data, you can provide a sample in the Add sample data section. If the rule finds matches in the sample data, a green Match label appears next to the input field.
Under Create keyword dictionary, you can add keywords to refine detection accuracy. For example, if you are scanning for a sixteen-digit Visa credit card number, you can add keywords like visa
, credit
, and card
.
Choose the number of characters before a match that the keyword must appear in. By default, keywords must be within 30 characters before a match.
Optionally, under Define rule target and action, edit the tags that you want to associate with events where the values match the rule. Datadog recommends using sensitive_data
and sensitive_data_category
tags, which can be used in searches, dashboards, and monitors. See Control access to logs with sensitive data for information on how to use tags to determine who can access logs that contain sensitive data.
For Set priority level, choose a value based on your business needs.
Click Update.
Navigate to the Sensitive Data Scanner using the Agent configuration page.
Hover over the scanning rule you want to edit and click the Edit (pencil) icon.
The Define match conditions section shows an explanation of the library scanning rule you chose along with examples of matched sensitive information.
Under Create keyword dictionary, you can add keywords to refine detection accuracy. For example, if you are scanning for a sixteen-digit Visa credit card number, you can add keywords like visa
, credit
, and card
.
Choose the number of characters before a match that the keyword must appear in. By default, keywords must be within 30 characters before a match.
Click Save.
Control access to logs with sensitive data
To control who can access logs containing sensitive data, use tags added by the Sensitive Data Scanner to build queries with role-based access control (RBAC). You can restrict access to specific individuals or teams until the data ages out after the retention period. See How to Set Up RBAC for Logs for more information.
To redact sensitive data contained in tags, you must remap the tag to an attribute and then redact the attribute. Uncheck Preserve source attribute
in the remapper processor so that the tag is not preserved during the remapping.
To remap the tag to an attribute:
- Navigate to your log pipeline.
- Click Add Processor.
- Select Remapper in the processor type dropdown menu.
- Name the processor.
- Select Tag key(s).
- Enter the tag key.
- Enter a name for the attribute the tag key is remapped to.
- Disable Preserve source attribute.
- Click Create.
To redact the attribute:
- Navigate to your scanning group.
- Click Add Scanning Rule.
- Check the library rules you want to use.
- Select Specific Attributes for Scan entire event or portion of it.
- Enter the name of the attribute you created earlier to specify that you want it scanned.
- Select the action you want when there’s a match.
- Optionally, add tags.
- Click Add Rules.
This functionality is not available for Sensitive Data Scanner using the Agent.
Cloud Storage scanning
Limited Availability
Scanning support for Amazon S3 buckets and RDS instances is in Limited Availability. To enroll, click Request Access.
Request AccessIf you have Sensitive Data Scanner enabled, you can catalog and classify sensitive data in your Amazon S3 buckets and RDS instances.
Sensitive Data Scanner scans for sensitive data by deploying Agentless scanners in your cloud environments. These scanning instances retrieve a list of all S3 buckets and RDS instances through Remote Configuration, and have set instructions to scan text files—such as CSVs and JSONs—and tables in every datastore over time. Sensitive Data Scanner leverages its entire rules library to find matches. When a match is found, the location of the match is sent to Datadog by the scanning instance. Data stores and their files are only read in your environment—no sensitive data is sent back to Datadog.
Along with displaying sensitive data matches, Cloud Storage surfaces any security issues detected by Cloud Security Management affecting the sensitive datastores. You can click any issue to continue triage and remediation within Cloud Security Management.
Out-of-the-box dashboard
When Sensitive Data Scanner is enabled, an out-of-the-box dashboard summarizing sensitive data findings is automatically installed in your account. To access this dashboard, go to Dashboards > Dashboards List and search for “Sensitive Data Scanner Overview”.
Disable Sensitive Data Scanner
To turn off Sensitive Data Scanner entirely, set the toggle to off for each Scanning Group so that they are disabled.
Further Reading
Documentation, liens et articles supplémentaires utiles: