How can you create a Cloud Monitoring dashboard to troubleshoot production service issues while differentiating between failures in your own service and those caused by a dependent Google Cloud service?