I'll get deeper later, but essentially this is how it works.
An agent is installed on each machine. This agent accesses the UCCAPI log file populated by the Skype for Business desktop application. This log file is rich with information on the status of the user's use of the Skype for Business client. The SfB Monitor agent currently filters for the <VQReportEvent> sections and enables that data for viewing on the desktop and also ships the data to an Azure tenant for viewing from an administrative console. Since the agent is always running, this data is available in real time. So, as long as a <VQReportEvent> event is generated, SfB Monitor will process it.
A little added benefit is since the agent looks at the UCCAPI log file, additional filters can be created to view events other than the <VQReportEvent>. For example, Exchange or Skype for Business logon failures can also be captured and much more.
In addition, since this application is client-side and does not derive any of its data from the Lync/Skype for Business server architecture, this can be applied into On-premises, Online, or Hybrid scenarios. This allows for a single datasource (Azure) for all client scenarios. This is not supposed to be a replacemtn for Call Quality Dashboard (CQD) or any 3rd Party solutions like UC Commander from EventZero or Prognosis from IR, it is supposed to be a new "user focused" way to gain real time visibility into call quality issues that users experience.