Queue Now monitor double pegging ACD calls offered/handled


Article ID: 51765 - Last Review: May 17, 2016

OVERVIEW 

After a call is handled in a Network ACD environment (where queues and agents exist on different PBXs), we typically get two SMDR records: one from the queuing gateway and another from the agent controller, and the two records arrive almost simultaneously. The prairieFyre summary logic relies primarily on the SMDR record from the queuing gateway to peg the queue (queue by period, queue totals etc...). However, we have seen cases where the 3330 does not report an SMDR record from the queuing gateway, and in these cases we rely on the SMDR record from the agent controller to peg the queue.


INFORMATION

The above summary logic works as expected in replay mode, esp. when doing a manual summarize or when the Maintenance service starts the summary for the previous day. However, in real-time mode, the timing of the records is critical for it to work properly. On our Inhouse, we have observed the SMDR record from the queuing GW arriving much later than that of the agent controller. The delay has been in the neighborhood of 35 to 50 seconds. Note that the timestamps and durations on these records match; it is just that they arrive late. Hence the reason it works properly in summary mode.

As an example: the following logs came from the collector service, right when the collector's TCP listener on port 1752 receives a block of raw SMDR data:

Record from the agent controller (Node 07, PBX sys ID 6)
Info 12/13/2012 11:22:46 AM Received SMDR data: /13 11:01:39  0000:21:03 X9999   0007 T6169400500 P286 606114     1325        006   6169400500           114                E0018104 C                                           

Record from the queuing GW (Node 01, PBX sys ID 1)
Info 12/13/2012 11:23:23 AM Received SMDR data: /13 11:01:39  0000:21:03 T8107   0857 P286 114 118 114            X9999           001   6169400500           7777               E0018104 A E0018582                                  

notice how the timestamps and durations on the above records are exactly the same, yet the records are 49 seconds apart.


RESOLUTION

The approach we took was to optionally not rely on SMDR data from the agent controller to peg queues. For this, we introduced a new configuration setting, IgnoreRecordsFromOtherMediasWhenPeggingQueues, set to false by default in order to maintain backwards compatibility. To enable this settings, here are the steps:

1. Browse to [Install Directory]CCM\Services\EnterpriseServer folder
2. Make a backup of the file prairieFyre.Services.EnterpriseServer.exe.config file.
3. Ensure the mentioned file is write enabled and open it in a text editor.
4. Look for the key IgnoreRecordsFromOtherMediasWhenPeggingQueues in the appSettings section of the file and set its value to 1 as indicated below. If not there, add it with a value of 1.

<add key="IgnoreRecordsFromOtherMediasWhenPeggingQueues" value="1"/>

5. Restart the prairieFyre Enterprise Server service.
6. Browse to [Install Directory]CCM\Services\DataProcessor folder
7. Make a backup of the file DataService.exe.config file.
8. Ensure the mentioned file is write enabled and open it in a text editor.
9. Look for the key IgnoreRecordsFromOtherMediasWhenPeggingQueues in the appSettings section of the file and set its value to 1 as indicated below. If not there, add it with a value of 1.

<add key="IgnoreRecordsFromOtherMediasWhenPeggingQueues" value="1"/>

10. Ensure the prairieFyre Data Processor service is stopped.

APPLIES TO

CCM/MiCC 6.0.2.0 and newer 

Keywords: queue now monitor offered handled double peg pegging


Posted Thursday, December 13, 2012 by amontpetit
http://micc.mitel.com/kb/KnowledgebaseArticle51765.aspx