DROPped calls when agents are waiting

[PManley]

Hello,

We have a VICI cluster set up with 10 dialer servers, 1 web server, 2 DB servers (Master and Slave) and recording/FTP server configured using VICI version VERSION: 2.14-760a BUILD: 200701-1525.

We have the system configured with 140 cps and 900 channels spread evenly across all 10 dialers. And on our cluster, we average about 200 agents logged in each day dialing on a blended campaign.

We are seeing cases where we have agents waiting for calls, but the DROP call counter is increasing during this time.

Can you please help me understand what could cause this? It seems that if we have agents waiting for calls, we should not be getting call DROPs.

Thanks and regards,

Peter Manley

[GenXOutsourcing]

When did this start? Always been this way? What changed? And, what are the server specs?

[PManley]

It has not always been this way and as far as we can tell any changes we made to the campaigns should not affect this. The server specs are as follows:

Master and Slave DB Servers:
Proliant DL380 Gen 9 with 128 GB RAM and DUAL Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz.

Dialer Servers:
Proliant DL320e Gen 8 v2 with 16 GB RAM and single Intel(R) Xeon(R) CPU E3-1271 v3 @ 3.60GHz

[GenXOutsourcing]

Ok, I will make a few assumptions here and please correct me if I am wrong.
200 agents, and I will assume you are load balancing across 10 servers.... so 20 agents per server, and 90 channels per server with 14cps per server.

You didnt say how many concurrent calls, or what dial level you are running. If you are using AMD, recordings, or both.

Now, what is the load on your servers. Your dialers are 4core, so at the bottom of the Reports page will show you the load. The DIALERs should not be exceeding 400. 400 is the same as 4.00 in top or htop. This would mean you are maxing out all 4 cores. Asterisk is happier at 50% or less, at 30% asterisk will love you all day long. SO your load should be between 200 and 400 maximum. When asterisk gets overloaded, it does crazy things........ not directing calls, having 3way calls between agents, losing calls, PDROPs, etc.

This is where I would start. Watch and see if this happens when you have less agents logged in. Does it start at a certain agent level?

When you have calls waiting and not sent to agents, note the IPs. Are they from 1 server, 2, or all of them.

But for me, without seeing your system, I am thinking overloaded asterisk.

[PManley]

@GenXOutsourcing,

We have approximately 300 current active calls. Dial Level is 2.5 with RATIO dialing method. We are not using AMD, but we are recording all calls.

The server load on the dialer servers ranges from 37 to 80. Never even close to 200. So, I don't think we're overloading Asterisk.

Our Hopper Loading Interval is 1 minute and the Min Hopper Level is set to 5000.