Considerations for configuration

First execution

If no only_logs_after value is provided, the module will only fetch the logs of the date of the execution.

Filtering

If the S3 bucket contains a long history of logs and its directory structure is organized by dates, it's possible to filter which logs will be read by Wazuh. There are multiple configuration options to do so:

• only_logs_after: Allows filtering logs produced after a given date. The date format must be YYYY-MMM-DD, for example, 2018-AUG-21 would filter logs produced after the 21st of August 2018 (that day included).

• aws_account_id: This option will only work on CloudTrail, VPC, and Config buckets. If you have logs from multiple accounts, you can filter which ones will be read by Wazuh. You can specify multiple ids separating them by commas.

• regions: This option will only work on CloudTrail, VPC and Config buckets, and Inspector service. If you have logs from multiple regions, you can filter which ones will be read by Wazuh. You can specify multiple regions separating them by commas. It is mandatory to specify the region when configuring an S3 bucket from an AWS GovCloud region (available GovCloud regions are us-gov-east-1 and us-gov-west-1).

• path: If you have your logs stored in a given path, it can be specified using this option. For example, to read logs stored in directory vpclogs/ the path vpclogs need to be specified. It can also be specified with / or \.

• aws_organization_id: This option will only work on CloudTrail buckets. If you have configured an organization, you need to specify the name of the AWS organization by using this parameter.

Older logs

The aws-s3 Wazuh module only looks for new logs in buckets based on the key of the last processed log object, which includes the datetime stamp. If older logs are loaded into the S3 bucket or the only_logs_after option date is set to a datetime earlier than previous executions of the module, the older log files will be ignored and not ingested into Wazuh.

On the other hand, the CloudWatch Logs module can process logs older than the first one processed. To do so, specify an older only_logs_after value, and the module will process all logs between the value set for only_logs_after and the first log executed without generating duplicate alerts.

# cd /var/ossec/wodles/aws
# ./aws-s3 -b 'wazuh-example-bucket' --reparse --only_logs_after '2021-Jun-10' --debug 2

Connection configuration for retries

Some calls to AWS services may fail when made in highly congested environments. The AWS pip dependencies client raises ClientError exceptions describing the errors. This kind of exception often needs repeating the call, without further handling. To help retry these calls, Boto3 provides Retries. This feature allows retrying client calls to AWS services when errors like ThrottlingException are experienced.

Users can customize two retry configurations.

• retry_mode: legacy, standard, and adaptive.

  • Legacy mode is the default retry mode. It sets the older version 1 for the retry handler. This includes:

    • Retry attempts for a limited number of errors/exceptions.

    • A default value of 5 for maximum call attempts.

  • Standard mode sets the updated version 2 for the retry handler. It includes:

    • Extended functionality over that found in the legacy mode where retry attempts apply to an expanded list of errors/exceptions.

    • A default value of 3 for maximum call attempts.

  • Adaptive mode is an experimental retry mode. It includes all the features of the standard mode. This mode offers flexibility in client-side retries. Retries adapt to the error/exception state response from an AWS service.

• max_attempts: The maximum number of attempts including the initial call. This configuration can override the default value set by the retry mode.

You can specify the retry configuration in the ~/.aws/config configuration file. The profile section must include the max_attempts, retry_mode, and region settings. It is important to use the same profile as the one you chose as your authentication method. If the authentication method doesn't have a profile, then the [Default] profile must include the configurations. In case the system doesn't have this file, the aws-s3 Wazuh module defines the following values by default:

• retry_mode=standard

• max_attempts=10.

The following example of a ~/.aws/config file sets retry parameters for the dev profile:

[profile dev]
region=us-east-1
max_attempts=5
retry_mode=standard

Configuring multiple services

Below there is an example of different services configuration:

<wodle name="aws-s3">
  <disabled>no</disabled>
  <interval>10m</interval>
  <run_on_start>yes</run_on_start>
  <skip_on_error>yes</skip_on_error>

  <!-- Inspector, two regions, and logs after January 2018 -->
  <service type="inspector">
    <aws_profile>default</aws_profile>
    <regions>us-east-1,us-east-2</regions>
    <only_logs_after>2018-JAN-01</only_logs_after>
  </service>

  <!-- GuardDuty, 'production' profile -->
  <bucket type="guardduty">
    <name>wazuh-aws-wodle</name>
    <path>guardduty</path>
    <aws_profile>production</aws_profile>
  </bucket>

  <!-- Config, 'default' profile -->
  <bucket type="config">
    <name>wazuh-aws-wodle</name>
    <path>config</path>
    <aws_profile>default</aws_profile>
  </bucket>

  <!-- KMS, 'dev' profile -->
  <bucket type="custom">
    <name>wazuh-aws-wodle</name>
    <path>kms_compress_encrypted</path>
    <aws_profile>dev</aws_profile>
  </bucket>

  <!-- CloudTrail, authentication with hardcoded keys (not recommended), without 'path' tag -->
  <bucket type="cloudtrail">
    <name>wazuh-cloudtrail</name>
    <access_key>XXXXXXXXXX</access_key>
    <secret_key>XXXXXXXXXX</secret_key>
  </bucket>

  <!-- CloudTrail, 'gov1' profile, and 'us-gov-east-1' GovCloud region -->
  <bucket type="cloudtrail">
    <name>wazuh-aws-wodle</name>
    <path>cloudtrail-govcloud</path>
    <regions>us-gov-east-1</regions>
    <aws_profile>gov1</aws_profile>
  </bucket>

  <!-- CloudTrail, 'gov2' profile, and 'us-gov-west-1' GovCloud region -->
  <bucket type="cloudtrail">
    <name>wazuh-aws-wodle</name>
    <path>cloudtrail-govcloud</path>
    <regions>us-gov-west-1</regions>
    <aws_profile>gov2</aws_profile>
  </bucket>

</wodle>

Using non-default AWS endpoints

<wodle name="aws-s3">
  <disabled>no</disabled>
  <interval>10m</interval>
  <run_on_start>yes</run_on_start>
  <skip_on_error>yes</skip_on_error>

  <bucket type="cloudtrail">
    <name>wazuh-cloudtrail</name>
    <aws_profile>default</aws_profile>
    <service_endpoint>https://bucket.xxxxxx.s3.us-east-2.vpce.amazonaws.com</service_endpoint>
  </bucket>

  <bucket type="cloudtrail">
    <name>wazuh-cloudtrail-2</name>
    <access_key>xxxxxx</access_key>
    <secret_key>xxxxxx</secret_key>
    <iam_role_arn>arn:aws:iam::xxxxxxxxxxx:role/wazuh-role</iam_role_arn>
    <sts_endpoint>xxxxxx.sts.us-east-2.vpce.amazonaws.com</sts_endpoint>
    <service_endpoint>https://bucket.xxxxxx.s3.us-east-2.vpce.amazonaws.com</service_endpoint>
  </bucket>

  <service type="cloudwatchlogs">
    <aws_profile>default</aws_profile>
    <regions>us-east-2</regions>
    <aws_log_groups>log_group_name</aws_log_groups>
    <service_endpoint>https://xxxxxx.logs.us-east-2.vpce.amazonaws.com</service_endpoint>
  </service>

</wodle>

<wodle name="aws-s3">
  <disabled>no</disabled>
  <interval>10m</interval>
  <run_on_start>yes</run_on_start>
  <skip_on_error>yes</skip_on_error>

  <service type="cloudwatchlogs">
    <aws_profile>default</aws_profile>
    <regions>us-east-2</regions>
    <aws_log_groups>log_group_name</aws_log_groups>
    <service_endpoint>logs-fips.us-east-2.amazonaws.com</service_endpoint>
  </service>

</wodle>