How to determine overall system health? RRS feed

  • Question


    I'm looking for a way to determine the overall OCS system health?  Something that is easy to understand for (Technical and not so Technical) Management.


    As for the background for this ... IT management is looking to determine how our pilot installation is working before "opening the taps" to thousands of users.  We currently have about 400 users in a pilot.  We're running the Enterprise Edition - Expanded.  There are nine servers:

    • Front-end server (which also hosts the Web Components server)
    • Web Conferencing server
    • A/V Conferencing server
    • SQL 2005 server (shared with other apps that need SQL)
    • Director
    • Archiving & CDR server
    • Quality of Experience (QoE) Monitoring Server
    • (internal access) Communicator Web Access server
    • (external access) Communicator Web Access server

    The DMZ isn't configured for external access yet (we need to modify our topology to allow for exposure of public IP on A/V Edge server) so we don't have that in place.


    Is there a counter(s), or something else, that can provide the average response time (perhaps for servers to fulfil requests), or some other metrics, that will provide us with the health our our environment?  Nothing too complicated - something that is easy to digest.  Thanks in advance, Joe.

    Monday, April 28, 2008 3:18 PM


All replies

  • I think SCOM is your only option


    Monday, May 19, 2008 9:06 PM
  • Thanks for the feedback...


    We've been looking at bring in Microsoft's Management Tools ... but we're not there yet.  Might be a year or more before they're integrated into our environment.


    In the mean time, I was looking for a fistful of counters (or something) that could give us "the big picture".

    Tuesday, May 20, 2008 12:43 PM
  • You can start out with normal Server counter to check basic System health, like CPU, Memory, Disk


    Tuesday, May 20, 2008 10:41 PM
  • Thanks for the feedback.  I appreciate the response.


    This is a good place to start, for generic server health ... the (generic server monitoring) counters are part of our regular performance monitoring.  It doesn't really touch on specific OCS performance/health, which is what I would like to monitor.


    We're piloting to 400 users and we use enterprise class servers.  The server health is pretty good (they are not being stressed) ... it's the OCS performance metrics that are important to us (management).  I'll do a little digging through the performance counters (there is a huge amount of them) and try to determine which are most beneficial when trying to gauge what the end-user experience is like.

    Wednesday, May 21, 2008 12:49 PM
  • Here's an article that might be useful to some:




    Anyone else have more information on the perfmon counters related to OCS?

    Thursday, June 5, 2008 9:11 PM
  • Thanks.  I recall seeing this article (one of many in the vast amount of articles on OCS) many moons ago.  Having actually read it now, I think this will suffice for a high level view of our system health.  Thanks to everyone for their input - I appreciate it...


    Friday, June 6, 2008 11:36 AM