vicidial.org

Hi Guys,

Trying to get to the bottom of an issue with calls being transferred to two people, two calls being dropped into same person etc.

Reading through previous posts, it seems like this is potentially an asterisk load issue that is causing it but looking for a bit of guidance on tools to check, places to look (eg mysql performance, any logging tools). Are there any screens I should monitor and will attaching to them cause issues with the processes that run on them?

I've picked an example and it seems that the load spikes on these servers at the time see image below just before 11AM. Are these spikes typical of what you would see on a standard vicidial install? Thanks in advance for your help.
https://ibb.co/cTcCvTR
https://ibb.co/qgXwdwV

VERSION: 2.14-721a
BUILD: 191015-1620
Asterisk 11.22.0-vici
Centos 6.10 AWS 4 core 8gb ram
6 Server Cluster, 1 Primary Web/SQL/Vici/Asterisk (No agents), 5 Vicidial/Asterisk
Approx 10-15 agents on each. Up to 20-30k calls a day.
MySQL v14.14 D5.1.73

We've seen this issue with a few clients recently, it ended up being a carrier issue where the inbound carrier in all cases.

I think this is more on the Vicidial/Asterisk side. The calls are fine coming in, its when they transfer (to another ingroup) that calls are going to two people. Sorry, I should have specified this properly in the first post.

I can't say I've run into that type of issue before. But, on overloaded systems all kinds of things can go wrong.

I had something similar when our DB wasn't strong enough to take the load.

Hi Guys,

We also have this type of issue, where an agent gets connected to two calls at same time but one is inbound and other is outbound in most of the cases.
I am not sure if this is due to overload, How can I check? (Sorry have no idea how to work with asterisk).

Thanks in advance.

Vidhi Shah wrote:Hi Guys,

We also have this type of issue, where an agent gets connected to two calls at same time but one is inbound and other is outbound in most of the cases.
I am not sure if this is due to overload, How can I check? (Sorry have no idea how to work with asterisk).

Thanks in advance.

Check the average server load on the servers (via uptime or htop). Particularly the DB server. DURING the event (not 10 minutes later, lol).

Also check to see if there are slow queries bogging down mysql and the mysql variables that track failures. See if those are incrementing higher during these events. You may be hitting an artificial mysql or OS configuration barrier/limit.

Ensure your Database server and other Vicidial servers are sharing a Gigabit local network that's clean and unobstructed (no dropped or delayed packets, preferably no firewall).

Log the events to begin a pattern recognition possibility. Number of Live agents (and on each server), live calls, ringing calls, inbound calls in queue, etc. Eventually you may find that it happens when specific conditions are met and then you may get an idea of where to search.

Alternately, if you can identify which call was routed incorrectly, you can follow the logs for that specific call to find where the logic failed and the call was sent to the wrong place. Every time we've chased this down, it's been due to an overloaded database server not able to respond quickly to a request and resulting in bad information being returned.