Lync Monitoring Server Summary (Clifs Notes)

The purpose of this post is to help identify the requirements to monitor your Lync Implementation holistically.  Real-Time montioring happens within the (optional) OpsMgr Managment Pack, and detailed reporting happens within the Lync Monitoring Server and the Reports deployed along with the Lync Monitoring Server Role.  In order to monitor your Lync Implementation fully, it is suggested to use both OpsMgr and the Lync Monitoring Server.

The Lync Monitoring Server simplifies the ability to gather data about the quality of your network media, in both Voice calls and A/V conferences. The Monitoring Server also collects call error records (CERs), which you can use to troubleshoot failed calls. The Monitoring Server collects usage material in the form of call detail records (CDRs) about various Lync Server components so that you can analyze ROI and plan the  capacity requirements of the Lync implementation. A single Monitoring Server is able to support up to 250,000 users if not collocated with an Archiving Server. If collocated, the Monitoring Server is able to support up to 100,000 users. Reporting is provided out of the box, but may not provide a holistic view of the entire platform the Lync Environment is deployed upon (SQL, IIS, OS).

Data Collection Agents  
·        Installed automatically on every Front End Server. The CDR agent intercepts SIP messages and sends data to the destination queue on the Monitoring Server. The QoE agent receives QoE data reports from endpoints through SIP SERVICE requests, and sends the data to the destination queue on the Monitoring Server or to third-party consumers using HTTP POST.
 
Monitoring Server  
·        Persists data received from Message Queuing to the SQL Server Monitoring Server database. This includes two parts: The CDR service and the QoE service. In order to collect data from a Registrar pool, you must use Topology Builder to associate the Monitoring Server with the Registrar pool.
 
Monitoring Server databases  
·        Run on SQL Server and store the captured data. There are separate databases for CDR and QoE information, but both always run on the same instance of SQL Server. Monitoring Server databases require the full edition of SQL Server. SQL Server Express is not supported.
 
Message Queuing  
·        Must run on each Monitoring Server and on each Front End Server that reports data to Monitoring Server. For each server, Message Queuing must be installed as Active Directory Domain Services (AD DS) integration mode so that the data can be delivered from Data Collection Agents to Monitoring Server.
 
(Optional) System Center Operations Manager management pack  
·        This is an optional component. The Call Reliability and Media Quality Monitoring of component of the System Center Operations Manager management pack uses Monitoring Server CDR and QoE data to generate near real-time alerts that show the health of call reliability and media quality. Near real-time alerts can be sent via messaging channels including SMTP, SMS, and IM in order to create alert notifications for proactive\preemptive monitoring.
 
(Optional) Monitoring Server Reports  
·        This is an optional component. This component contains out-of-the-box reports providing usage, call diagnostic information and media quality information, based on CDR and QoE data stored in the CDR and QoE databases. The reports are generated with SQL Server Reporting Services.
 
Richer Reporting, including Dashboard View
·        Monitoring Server takes advantage of Microsoft SQL Server Reporting Services to provide richer reporting on system usage for ROI analysis, voice quality for diagnostics scenarios, and IP phone and media device asset management.
·        A dashboard presents an aggregation of these reports, proactively informing administrators about usage trends and the health of the deployment.
 
Comprehensive Data for Usage Tracking and Voice Quality Diagnostics
·        The call detail recording (CDR) and Quality of Experience (QoE) databases have been updated to include new usage and diagnostic data for all new Lync Server 2010 features, conferencing, registration, and device diagnostics. New database schemas incorporate these changes.
 
New Management Features
·        Monitoring Server now uses Lync Server Management Shell cmdlets for all administration and management tasks. Lync Server provides centralized deployment for Monitoring Server from Topology Builder and the Deployment Wizard.
 
Optimized Infrastructure
·        The Monitoring Server infrastructure has been optimized to improve reliability and maintainability.
 
Scaling and Location of Monitoring Server
·        When you use the recommended hardware configuration, and you collocate the Monitoring Server and Monitoring database on the same computer, a single Monitoring Server can serve up to 250,000 users.
·        If you have multiple pools that total less than 250,000 users, it is recommend that you associate all of these pools with a single Monitoring Server to simplify management.
·        Alternatively, if you have pools located in different physical locations, it may make more sense to deploy a Monitoring Server at each location.
 
Monitoring Database Performance
·        For optimal performance, best practice recommends that you put the following files on four physical disks:
o   System file and Message Queuing file on the same physical disk
o   QoE database data file and CDR database data file on the same physical disk
o   QoE database log file
o   CDR database log file
·        If you collocate the Monitoring Server databases with other databases on the same server, you should run the Monitoring Server databases in a separate instance from other databases.
·        It is recommended to put the Monitoring Server database data files and log files on separate physical disks for optimal performance.
·        You should carefully evaluate performance impacts before deciding to collocate the Monitoring Server databases with other databases.
 
Monitoring Database Size
·        Based on the Lync Server user model
o   The CDR database grows 31.5 KB per user per day
o   The QoE database grows 28 KB per user per day
·        You can use this formula,  (Database size = (DB growth per user per day) * (Number of users) * (Number of days) along with the knowledge of your available database disk space, to help decide how many days of data to keep on your database (the default is 60 days).
o   For example, 60 days of data in the CDR database for 50,000 users would be 31.5*50000*60 for a total of 90 GB. If your organization’s Lync Server differs significantly from the user model, adjust the daily database growth estimate.
 
Report Performance
Reporting is a factor affected by performance.
·        The provided standard set of reports are designed to work in most scenarios, but if you need reports on very large data, such as a QoE report on 10 million calls, an offline reporting solution might be more appropriate.
·        It is recommend that you query reports during a non-busy time in order to avoid resource conflicts with data insertion.
·        If your Monitoring database size is larger than the database server’s physical memory while querying, Monitoring Server report performance can be affected.