One trick I've noticed that nobody is doing for outbound limit, is using "execute_on_originate" to do the limit there on each channel. You can specify it in the local variables per channel.
The following would do it for each B-Leg, w/o messing associating the coun't to your a-leg session:
bridge:[execute_on_originate=limit hash foo gwa 10]sofia/gateway/gwa/$1|[execute_on_originate=limit hash foo gwb 10]sofia/gateway/gwb/$1 ... and so on...
I use this all the time in a more extended way. I use execute_on_originate to call my own dialplan app, but you could use it to invoke a lua script as well on the session of the B-Leg, before the call is originated. If you just return from your dialplan app, the b-leg will continue, but if you hangup the channel, then it will look like the b-leg failed with the hangup cause you specify, and will get passed back as if the far end returned that cause.
--Dave
From: Muhammad Naseer Bhatti [mailto:nbhatti@gmail.com]
To: FreeSWITCH Users Help [mailto:freeswitch-users@lists.freeswitch.org]
Sent: Tue, 29 May 2012 14:12:01 -0700
Subject: Re: [Freeswitch-users] Gateway failover with LUA
They are two different gateways with two different names, but the same
IP address yet the port is different. Possible that limit subsystem is
imposing the limit on the IP address itself and not able to
distinguish between the different ports.