Debugging why Asterisk Manager is slow to Originate calls

I am currently running Asterisk and I’m using Centreon for configuration by way of the Fully Automated Nagios installation.

We have a pair of PRIs in a dedicated Dell 1800. Calls are logged to MySQL cdr database.

To make a call in Asterisk I am sending commands like this from a php script that runs when a phone number is clicked:
Action: Originate
Channel: SIP/$deskPhoneExt
Context: from-click-to-dial
Exten: $numberToCall
Priority: 1
Callerid: Click to Dial <$callerIdNumber>

The script exits indicating the call was passed to Asterisk. But sometimes through the day it may be up to a couple minutes later that the phone actually rings. It happens in short spurts of up to 8 minutes a couple times a day but with 50 people using the phone I’m getting quite a bit of negative feedback. The issue seems to be becoming more frequent.

The quality of the calls seems to be fine. The server load average doesn’t seem remarkable, last time it was under .20 Because the script is finishing it doesn’t seem to be related to bandwidth. It doesn’t seem to be caused by anything running in cron or another background process as most of those run at night.

I tried to connect gdb to it for a backtrace but that freezes calls while running :-/ I’ll have to do more exparamenting to get gdb to connect and see if disconnecting it drops calls so that I could try loading it with it on in the morning and just disconnecting later if things are slow. UPDATE: I see it only freezes it while the command is running, I’ll run “thread apply all bt” and exit quickly the next time it happens so I can provide a backtrace.

The issue doesn’t seem to be related to the number of currently connected calls.

Any help on things I could try would be greatly appreciated.


I am having the same problem,

does anyone else have faced with this problem? Anyone any ideas how to solve this?