Skip to main content

Cradle-to-Grave Call History

Posted by KevinH on Thu, 03/30/2017

One request we get fairly frequently from clients that have been brought over to Thirdlane is to have a user-friendly Call History section.
When a call comes in, goes to a hunt group, no one answers and it goes to voicemail the office manager wants to know why it went to voicemail because they don't see that the call went through the hunt group. This is just one example but being able to see the objects between Answer() and Hangup() would make the system better understood by the end user and really help sell it.

In one instance we wrote a cradle-to-grave CDR reporting tool for a client because we were going to lose their business over it. Supporting it is painful and it not being a part of the standard GUI feels unpolished so this isn't a scalable solution we can roll out to multiple clients.


Submitted by eeman on Tue, 04/11/2017 Permalink

Kevin, this would be something limited to STE in my opinion due to the load this would drive up onto a server. The goal of MTE is scalability and nobody wants to be told you can only have 200 extensions on an MTE box due to the 100x disc I/O implemented by rolling out Channel Event Logging (CEL) within asterisk. The tool you have seen in other GUI is CEL. This is what switchvox uses when showing you those details. There is a reason why Switchvox is only single tenant flavor and tells you shit like 'this box can have 400 extensions and only 75 concurrent calls'. A call center with 400 agents and only 75 concurrent calls is a call center that needs to fire 325 agents pretty damn quick if you get my meaning. Your're literally paying 325 agents to _not_ answer the phone. These limits exist because of the increase load they are taxing the system with performing these other tasks.

MTE is a dichotomy in that scalability and features often find themselves at opposite ends of the spectrum. Many are pushing MTE toward its maximums without having to have 20 racks full of 1U servers just to handle thousand calls. MTE needs to gear toward attaining 5000 concurrent calls (as long as RTP range is 10000 - 20000 then 5000 is your call limit). 5000 concurrent calls is not possible if the system has to perform a hundred different writes for every single call. The FULL log by itself is a hindrance to this goal because it is constantly enabled and every call generates several hundred lines of data. At the same time if the school of thought for MTE remains to keep offloading services onto other hardware, sales will plummet when faced with huge startup costs. Not only will you have to purchase additional servers for all the offloaded tasks, but you will have to pay for their rackspace, electricity, cooling, etc.

STE on the other hand, I fully support the ability to use CEL and view its content. CEL is an all-or-nothing logging system.