I am getting the following alert in SCOM related to my Lync 2010 AV conferencing server..

Perf Object Name: LS:MEDIA - 01 - Planning
Perf Counter Name: MEDIA - 009 - Number of occasions conference processing is delayed significantly
Perf Instance Name: AVMCUSvc.exe
Perf Counter Value: 53

The counter value is now increased and become 21000.

I have read some blog posts and articles which relates this counter to high CPU usage and performance issue with AV conferencing server.

I am not able to find any type or performance issues with my AV Conf servers and I have not received any user calls related to issues with conferences but the counter keeps getting higher and higher until I reboot the server.

Can anyone explain the counter and let me know how it can be fixed, if this is an issue?

From event log, I see lot of following errors, but not sure if it's related..

Log Name:      Lync Server
Source:        LS Audio-Video Conferencing Server
Date:          07/16/2015 9:30:21 AM
Event ID:      32010
Task Category: (1017)
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      LYNC01.Domain.com
Description:
A user attempted to connect to a conference, but an internal exception occurred while processing the request.  The connection attempt was rejected.
 
Conference URI: sip:User1@Domain.com;gruu;opaque=app:conf:focus:id:2tfncy63, User URI: sip:User2@Domain.com, Exception: Error in the application.

Hi,

Do you have the last cumulative update installed in the server? Also, do you have some edge in the topology?

https://support.microsoft.com/en-us/kb/2493736

https://support.microsoft.com/en-us/kb/2539646

Do you have a system baseline to compare the load difference?

Thanks / rgds,

@Thiago, yes Lync server is updated with latest CU.

@Desmond, We recently installed the SCOM and started monitoring Lync servers, I am in process to fine tune the alerts now.. Are you saying 'Number of occasions conference processing is delayed significantly' is just a performance indicator which we need to adjust the threshold based on our servers load?

I feel this is like a incremental counter, when I reboot the server it resets and start from 0 and gradually grows to 25K. It continue to grow until I reboot the server again. As per the SCOM monitor the threshold is 30 and considering that 25K count is really huge..

Hi -Edatha- ,

 

Do you have the latest MP imported on your SCOM ? (the latest one should have 4.0.7577.203 I think)

This perf counter increments when the conferencing server is not able to process the audio in real time. An example of this is when the server is overloaded and is not able to keep up with audio processing real time (or if there are virtualized servers with high load and low machine config), or if there is a process which is running with high-pri real time threads and starving the AVMCU of the real time CPU it needs.

Since you did not receive any user calls related to issues with conferences, I suggest you can observe this issue for a period of time and then consider remove this monitoring alert from SCOM.

 

Best regards,

Eric

Hi -Edatha-,

Im writing to follow up on this thread. Do you have any updates on this issue? Please post back at your convenience.

 

Best regards,

Eric

Thank You Eric, We have updated MP for Lync 2010 on SCOM. Is there any way we can identify the conference calls which are delayed? if not as you said we can observe the server performance for few days and think of disabling this alert.