Back to Checks
aws

Glue Data Catalog is not publicly accessible via its resource policy

glue_data_catalogs_not_publicly_accessible
Service: glueSeverity: high
OfficialBuilt-inVerified
Categories:internet-exposed
Resource Type:Other

AWS Glue Data Catalog resource policies are assessed for configurations that expose the catalog to anyone, such as Principal: *, broad resource scopes, or permissive conditions.

The finding highlights catalogs made public through overly permissive resource-based access.

Prowler logoProwlerprowler-provider-aws

Risk

Public catalog access lets unauthorized actors enumerate schemas, S3 locations, and connection metadata, weakening confidentiality. If writes are exposed, attackers can alter databases/tables, corrupt lineage, and disrupt jobs and queries, harming integrity and availability, and enabling lateral movement to data stores.

Run this check with Prowler

Run in Prowler Cloud
Run
prowler aws --checks glue_data_catalogs_not_publicly_accessible

Recommendation

Enforce least privilege on catalog resource policies:

  • Avoid Principal: * and wildcards
  • Grant only required actions to explicit principals
  • Prefer identity-based access or Lake Formation for sharing
  • Limit scope with precise ARNs/conditions and monitor changes for defense in depth

Remediation

CLI

aws glue delete-resource-policy

Terraform

resource "aws_glue_resource_policy" "<example_resource_name>" {
  policy = jsonencode({
    Version   = "2012-10-17",
    Statement = [
      {
        Effect    = "Allow",
        Principal = { AWS = "arn:aws:iam::<ACCOUNT_ID>:root" } # Critical: restricts to your account, removing any public (*) access
        Action    = "glue:*",
        Resource  = "arn:aws:glue:<REGION>:<ACCOUNT_ID>:catalog"
      }
    ]
  })
}

Other

1Sign in to the AWS Console and open the Glue service
2In the left menu, click Settings
3Under Data catalog settings > Permissions, click Edit resource policy
4Remove any statement that has Principal set to * (public) or AWS: "*"; or delete the entire policy
5Click Save

Check Code

from prowler.lib.check.models import Check, Check_Report_AWS
from prowler.providers.aws.services.glue.glue_client import glue_client
from prowler.providers.aws.services.iam.lib.policy import is_policy_public


class glue_data_catalogs_not_publicly_accessible(Check):
    def execute(self):
        findings = []
        for data_catalog in glue_client.data_catalogs.values():
            if data_catalog.policy is None:
                continue
            report = Check_Report_AWS(metadata=self.metadata(), resource=data_catalog)
            report.resource_id = glue_client.audited_account
            report.resource_arn = glue_client._get_data_catalog_arn_template(
                data_catalog.region
            )
            report.status = "PASS"
            report.status_extended = "Glue Data Catalog is not publicly accessible."
            if is_policy_public(
                data_catalog.policy,
                glue_client.audited_account,
            ):
                report.status = "FAIL"
                report.status_extended = "Glue Data Catalog is publicly accessible due to its resource policy."

            findings.append(report)

        return findings

References