Ringtime much lower on remote server

General Help
1

Greetings.

We have got 2 PBX, with no overlapping queues nor agents, and many of them configured as remote agents.

Recently, we enlarged all this queues to 3 times their original size (also being the remote agents 3 times more in every queue) and we have detected that, for any call that goes to the remote server due to local agents being busy, those calls rings for just 2-4 seconds, instead of the 15 defined by queue.

General settings default ringtime is defined at 15 too, and extensions ringtime to default.

Due to my lack of knowledge about PBX, im not able to track it further, any suggestion why is this happening or how could I diagnose it?

We have tested, creating a new queue, being just 1 remote agent in that queue, with no locals, that calls rings for about 9-13 secs, whenever we add the bulk, it gets lowered to 2-7 secs.

Machine overall perfomance is good (no more than 35% CPU 70% RAM at peak times).

Thanks for your help.

2

I’ll try to add some more info and background scenario.

FreePBX is barely new to me, I was forced to inherit the platform, and I’m learning day by day.

This is generated because we have 2 instances of PBX in different geographics zones, we overflow our local extensions to remote ones via weight, having them all in the local queue.

My previous mate, a VoIP specialist, was trying to get on running distributed states between instances so, as far as all the agents, remote and local, are on local queues, FreePBX could knew remote agents state and not dial busy remote agents. He wasn’t able to stop a Java memory leak when he rollbacked it before he left us.

So the scenario now running is that, in general, we got more remote agents than local ones, so the call have a slight dely finding an available agent, which is not a critical problem, yet, when it finds him, instead of ringing the 15 seconds defined by queue, it just makes 2-4 seconds ring.

I don’t find myself able to fix or get ready what he wasn’t, state distribution, so what I am looking for is how could we troubleshoot this.

I’ve been checking Asterisk Verbose to find any information I could paste here to add more information, but as far as I can read the verbose, this problem isn’t detected here, I found it via user complain, and using asternic call stats, where we can see that ringtime is 15000 ms, but the time between the previous event and that, in queue time, is no longer than 2 seconds. Having a remote sessions with our afected operators confirmed that their softphone (X-lite) didn’t hold the call for any longer.

Thanks for your support.

3

Asterisk (Ver. 1.8.23.0)

FreePBX Version 2.10.0

I don’t know if its a weird problem which its hard to diagnose, or I still lack useful info for you.

I’m at your service in anything else you could need in order to help me out.

Regards.

4

Could it be possible that the 15secs Agent Timeout window is opened at the first check, and not when it rings an available user?

I mean, lets suppose that our instance checks the status of 12 remote agents, getting busy from them all, and spending near 1 sec per check, would that mean that it will have only 3 secs left to ring on the 13th agent, who is free?

This is hypothetical, since for the same queue and agents logged in, we got ringtimes from 2 to 12 secs, having 15 secs agent timeout.

If it works by default this way, is there anything we could made to open the 15sec window when a free agent is detected?

Regards.