aws-sagemaker-executes-as-mlflow
An MLflow Tracking Server executes as its RoleArn.
match (record)
{
"field": "RoleArn",
"resource_type": "AWS::SageMaker::MlflowTrackingServer"
}
where
RoleArn is non-empty
emit
| source type | ApplicationPlatform |
|---|---|
| target type | MachineIdentity |
| source | <SageMaker MlflowTrackingServer> |
| target | <RoleArn> |
| conditions | service_state |
Narrative
SageMaker MLflow tracking server {server.name} runs as {role.name}, giving it access to S3 model artifacts and other resources the role permits.
Raw rule rules/explicit/aws/sagemaker.yaml
id: aws-sagemaker-executes-as-mlflow
emits: ExecutesAs
description: An MLflow Tracking Server executes as its RoleArn.
match_record:
resource_type: AWS::SageMaker::MlflowTrackingServer
field: RoleArn
where:
- RoleArn is non-empty
emit:
source_type: ApplicationPlatform
target_type: MachineIdentity
source: <SageMaker MlflowTrackingServer>
target: <RoleArn>
state: CONDITIONAL
conditions:
- service_state
permissions: []
false_positive_note: Only relevant when the MLflow server is Created/Running. Role used to read/write
model artifacts to S3. Low direct escalation value unless the role has broad S3 or IAM permissions.
narrative: SageMaker MLflow tracking server {server.name} runs as {role.name}, giving it access to S3
model artifacts and other resources the role permits.