vicidial.org

We are running Vicidial 3 box cluster (mysql, web + asterisk11, asterisk11) all scratch install on Centos6 dedicated hosted servers 100mb internet.

When first built cluster, we had only 2 boxes (mysql and webserver+asterisk) and all worked perfectly. As we used the system more and more, we started receiving an excessive amount of CONGESTION and CANCELs no matter which provider we used as carrier (have tried 6 or 7 with same results). We started at a dial level of 2 and during the following weeks needed to increase to level 6 or 8 for agents to receive calls. With 15 agents logged in we were dialing 80-100 lines. We added another Asterisk only box, and the dial level dropped back to 2 or 3 with less CONGESTION and CANCELs, however now we are faced with the same issue of increasing dial limit to 8 or 10 sometimes dialing 130-140 lines. Also we have noticed that the asterisk CLI is showing BROKEN PIPE errors which never showed before in Asterisk11. Sometimes we can start the day at dial level 2 but at the end of day are at level 10.

Any help or suggestions would be appreciated

The broken pipe errors can be ignored. As to your other problems, have you tried using another carrier?

What is your calls per second set to for your dialers?

Have you looked at the Carrier Log Report to see what the SIP codes and messages are on these non-connected calls?

Turn on Enhanced Disconnect Logging. (Admin->System Settings)

Stop calling "NA" as a test to see if you are merely collecting CONGESTION and CANCEL calls from previous runs. It's fairly common to have these build up and eventually fill your calling lists.

If removing "NA" from dialable statuses improves your results, you need fresh leads. Also consider "Call Count Limit" of 8 instead of 0 (Campaign Detail View), as this can help in the long run to get some use of the NAs that may answer and eventually dump them before they get too overused.