Playbook 1: S3 ACL Public Access

Introduction

This playbook describes how to configure Dow Jones Hammer to identify S3 buckets that are publicly accessible because of the ACL settings.

1. Issue Identification

Dow Jones Hammer investigates S3 buckets to identify Grantee parameter values AllUsers and AuthenticatedUsers.

When Dow Jones Hammer detects an issue, it writes the issue to the designated DynamoDB table.

According to the Dow Jones Hammer architecture, the issue identification functionality uses two Lambda functions. The table lists the Python modules that implement this functionality:

Designation Path
Initialization hammer/identification/lambdas/s3-acl-issues-identification/initiate_to_desc_s3_bucket_acl.py
Identification hammer/identification/lambdas/s3-acl-issues-identification/describe_s3_bucket_acl.py

2. Issue Reporting

You can configure automatic reporting of cases when Dow Jones Hammer identifies an issue of this type. Dow Jones Hammer supports integration with JIRA and Slack. These types of reporting are independent from one another and you can turn them on/off in the Dow Jones Hammer configuration.

Thus, in case you have turned on the reporting functionality for this issue and configured corresponding integrations, Dow Jones Hammer, as defined in the configuration, can:

  • raise a JIRA ticket and assign it to a specific person in your organization;
  • send the issue notification to the Slack channel or directly to a Slack user.

Additionally Dow Jones Hammer tries to detect person to report issue to by examining owner tag on affected S3 bucket. In case when such tag exists and is valid JIRA/Slack user:

  • for JIRA: jira_owner parameter from ticket_owners.json is ignored and discovered owner is used instead as a JIRA assignee;
  • for Slack: discovered owner is used in addition to slack_owner value from ticket_owners.json.

This Python module implements the issue reporting functionality:

hammer/reporting-remediation/reporting/create_s3bucket_acl_tickets.py

3. Issue Remediation

3.1 Automatic

To reduce the workload of your DevOps engineers and mitigate the threats stemming from this issue, you can configure automatic remediation of issues. It means that in case Dow Jones Hammer has detected and reported an issue, but the assignee of the report has not remediated the issue within a timeframe specified in the configuration, the Dow Jones Hammer remediation job will adjust S3 bucket ACL to eliminate this vulnerability.

In this specific case, Dow Jones Hammer will replace the current S3 ACL with the private Canned ACL (a set of AWS predefined grants that only provides the bucket’s owner with FULL_CONTROL and restricts access for everyone else).

Dow Jones Hammer will save the pre-remediation S3 ACL configuration to a S3 backup bucket. You can use this configuration to rollback the automatic remediation manually, if necessary.

This Python module implements the issue remediation functionality:

hammer/reporting-remediation/remediation/clean_s3bucket_acl_tickets.py

3.2 Manual

To retain full control on the remediation functionality you can disable automatic remediation in config.json and launch it manually:

  1. Login to Dow Jones Hammer reporting and remediation EC2 via SSH with centos user and ssh key you created during deployment: ssh -l centos -i <private_key> <EC2_IP_Address>
  2. Become root user: sudo su -
  3. Change directory to Dow Jones Hammer sources: cd /hammer-correlation-engine
  4. Launch Dow Jones Hammer remediation script: python3.6 -m remediation.clean_s3bucket_acl_permissions
  5. Confirm or refuse remediation of each issue separately

4. Setup Instructions For This Issue

To configure the detection, reporting, and remediation of this issue, you should edit the following sections of the Dow Jones Hammer configuration files:

4.1. The config.json File

The config.json file is the main configuration file for Dow Jones Hammer that is available at deployment/terraform/accounts/sample/config/config.json. To identify, report, and remediate issues of this type, you should add the following parameters in the s3_bucket_acl section of the config.json file:

Parameter Name Description Default Value
enabled Toggles issue detection for this issue true
ddb.table_name Name of the DynamoDB table where Dow Jones Hammer will store the identified issues of this type hammer-s3-public-bucket-acl
accounts Optional comma-separated list of accounts to check and report for this issue type aws.accounts from config.json
remediation_accounts Optional comma-separated list of accounts to remediate this issue type aws.accounts from config.json
reporting Toggle Dow Jones Hammer reporting functionality for this issue type false
remediation Toggle Dow Jones Hammer automatic remediation functionality for this issue type false
remediation_retention_period The amount of days to pass between issue detection and its automatic remediation. The value 0 denotes that Dow Jones Hammer will remediate the issue at the next remediation job run. 0

Sample config.json section:

"s3_bucket_acl": {
    "enabled": "true",	
    "ddb.table_name": "hammer-s3-public-bucket-acl",
    "reporting": false,
    "remediation": false,
    "remediation_retention_period": 0
}

4.2. The whitelist.json File

You can define exceptions to the general automatic remediation settings for specific S3 buckets. To configure such exceptions, you should edit the s3_bucket_acl section of the whitelist.json configuration file as follows:

Parameter Key Parameter Value(s)
AWS Account ID S3 bucket name(s)

Sample whitelist.json section:

"s3_bucket_acl": {
    "123456789012": ["test-iam-access-key-user", "AKIAIGZY37NNDDWXQNOA"]
}	

4.3. The ticket_owners.json File

You should use the ticket_owners.json file to configure the integration of Dow Jones Hammer with JIRA and/or Slack for the issue reporting purposes.

You can configure these parameters for specific AWS accounts and globally. Account-specific settings precede the global settings in the ticket_owners.json configuration file.

Check the following table for parameters:

Parameter Name Description Sample Value
jira_project The name of the JIRA project where Dow Jones Hammer will create the issue AWSSEC
jira_owner The name of the JIRA user to whom Dow Jones Hammer will assign the issue Support-Cloud
jira_parent_ticket The JIRA ticket to which Dow Jones Hammer will link the new ticket it creates AWSSEC-1234
slack_owner Name(s) of the Slack channels (prefixed by #) and/or Slack users that will receive issue reports from Dow Jones Hammer ["#devops-channel", "bob"]

Sample ticket_owners.json section:

Account-specific settings:

{
    "account": {
        "123456789012": {
            "jira_project": "",
            "jira_owner": "Support-Cloud",
            "jira_parent_ticket": "",
            "slack_owner": ""
        }
    },
    "jira_project": "AWSSEC",
    "jira_owner": "Support-General",
    "jira_parent_ticket": "AWSSEC-1234",
    "slack_owner": ["#devops-channel", "bob"]
}

5. Logging

Dow Jones Hammer uses CloudWatch Logs for logging purposes.

Dow Jones Hammer automatically sets up CloudWatch Log Groups and Log Streams for this issue when you deploy Dow Jones Hammer.

5.1. Issue Identification Logging

Dow Jones Hammer issue identification functionality uses two Lambda functions:

  • Initialization: this Lambda function selects slave accounts to check for this issue as designated in the Dow Jones Hammer configuration files and triggers the check.
  • Identification: this Lambda function identifies this issue for each account/region selected at the previous step.

You can see the logs for each of these Lambda functions in the following Log Groups:

Lambda Function CloudWatch Log Group Name
Initialization /aws/lambda/hammer-initiate-s3-acl
Identification /aws/lambda/hammer-describe-s3-acl

5.2. Issue Reporting/Remediation Logging

Dow Jones Hammer issue reporting/remediation functionality uses /aws/ec2/hammer-reporting-remediation CloudWatch Log Group for logging. The Log Group contains issue-specific Log Streams named as follows:

Designation CloudWatch Log Stream Name
Reporting reporting.create_s3_acl_tickets
Remediation remediation.clean_s3_acl

5.3. Slack Reports

In case you have enabled Dow Jones Hammer and Slack integration, Dow Jones Hammer sends notifications about issue identification and remediation to the designated Slack channel and/or recipient(s).

Check ticket_owners.json configuration for further guidance.

5.4. Using CloudWatch Logs for Dow Jones Hammer

To access Dow Jones Hammer logs, proceed as follows:

  1. Open AWS Management Console.
  2. Select CloudWatch service.
  3. Select Logs from the CloudWatch sidebar.
  4. Select the log group you want to explore. The log group will open.
  5. Select the log stream you want to explore.

Check CloudWatch Logs documentation for further guidance.

6. Issue specific details in DynamoDB

Dow Jones Hammer stores various issue specific details in DynamoDB as a map under issue_details key. You can use it to create your own reporting modules.

Key Type Description Example
owner string Bucket owner’s display name (available not for all regions) test-user
public_acls map Permissions grouped by Amazon S3 Predefined Groups {"AllUsers": ["WRITE", "READ"]}
tags map Tags associated with bucket {"Name": "TestBucket", "service": "archive"}