Detections
Analytics

DKER-EE-001590 - Docker Enterprise must alert the ISSO and SA (at a minimum) in the event of an audit processing failure.

Information

The Universal Control Plane (UCP) and Docker Trusted Registry (DTR) components of Docker Enterprise provide audit record generation capabilities. Audit logs capture all HTTP actions for the following endpoints: Kubernetes API, Swarm API and UCP API. The following UCP API endpoints are excluded from audit logging (where '*' designates a wildcard of exclusions): '/_ping', '/ca', '/auth', '/trustedregistryca', '/kubeauth', '/metrics', '/info', '/version*', '/debug', '/openid_keys', '/apidocs', 'kubernetesdocs' and '/manage'. Audit log verbosity can be set to one of the following levels: 'none', 'metadata', or 'request'. To meet the requirements of this control, the 'request' verbosity level must be configured in UCP.

The data captured at each level for UCP and the eNZI authentication and authorization backplane is described below:

'none': audit logging is disabled

'metadata':
 - method and API endpoint for the request
 - UCP user which made the request
 - response status (success/failure)
 - timestamp of the call
 - object ID of created/updated resource (for create/update calls)
 - license key
 - remote address

'request': includes all fields from the 'metadata' level, as well as the request payload

DTR audits all events associated with repository activities. Events are considered as follows: create, get, delete, update, send, fail, and scan. The following types are associated with the defined audit events: repository, tag, blob, manifest, webhook, uri, promotion, push mirroring, poll mirroring, garbage collector, system.

All audit logs generated by UCP and DTR can be forwarded to a remote log aggregation system by configuring an appropriate log driver plugin on all Docker Engine - Enterprise nodes in a cluster.

The Docker Engine - Enterprise component of Docker Enterprise relies on the underlying host operating system's auditing capabilities. By default, the host OS is not configured to audit Docker Engine - Enterprise.

Solution

via CLI:

Linux: As a trusted user on the host operating system, open the /etc/docker/daemon.json file for editing. If the file doesn't exist, it must be created.

Set the 'log-driver' property to one of the following: 'syslog', 'awslogs', 'splunk', 'gcplogs', 'logentries' or '<plugin>' (where <plugin> is the naming of a third-party Docker logging driver plugin). Configure the 'log-opts' object as required by the selected 'log-driver'.

Save the file. Restart the docker daemon.

Work with the SIEM administrator to configure an alert when no audit data is received from Docker.

See Also

https://dl.dod.cyber.mil/wp-content/uploads/stigs/zip/U_Docker_Enterprise_2-x_Linux-Unix_V2R2_STIG.zip

Item Details

Category: AUDIT AND ACCOUNTABILITY

References: 800-53|AU-5a., CAT|III, CCI|CCI-000139, Rule-ID|SV-235787r960912_rule, STIG-ID|DKER-EE-001590, STIG-Legacy|SV-104745, STIG-Legacy|V-95607, Vuln-ID|V-235787

Plugin: Unix

Control ID: 0df13a3e78c6f5784fad4ce9a7bd02f9824117403ab0233ffa5263e0ba2bdd91