Failed call handlers in Unity connection

Unanswered Question
Apr 20th, 2012

Hello. If I have a scenario where I try to transfer to a call handler in unity connection and the call handler fails (ex. Not reachable, busy, etc) is there rules to route this to other destinations? For example transfer to another call handler or extension? I have unity connection 8.6.


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Jaime Valencia Fri, 04/20/2012 - 20:54

If you're using the transfer option to release to switch control is not any longer on CUC, whatever behavior you might have configured on CUCM for the DN will be used for the call.



if this helps, please rate

Phil Bradley Sat, 04/21/2012 - 08:11

Hey Java,

The call handler actually gets passed to a route pattern for a sip trunk. I did not see any option on the route pattern or sip trunk where I could forward on failures or busy like on a DN.


Jaime Valencia Sat, 04/21/2012 - 17:27

Because there aren't any for that behavior if you're using a route pattern, on that scenario you're depending on what the destination phone has configured and CUCM won't pull back the call.

Like wise, if the call fails because of wrong sip trunk settings or because the other end failed to get the call or whatever other reason, that's it, it'll fail and you cannot do any more handling.

The other end would need to have configured to send the call back to CUCM or CUC if you want to do something else.



if this helps, please rate

Phil Bradley Sat, 04/21/2012 - 18:41

Ok, is there another way to send a call to a SIP trunk besides a route pattern and detect if the link is down? I was hoping that CUCM could at least ping the SIP trunk and if it is unreachable be able to route to another location. My scenario is that I have a Asterisk based automated system and if for some reason the server has went down I would like to route to a live agent (a contact center cti route point for example).


Jaime Valencia Sat, 04/21/2012 - 20:52

No, if you want to send a call via a sip trunk you need a route pattern.

And no, CUCM doesn't ping any SIP trunks to detect if they're alive or not, you need to get a response from it to try routing another way, and you can only route via route groups and route lists once you're using a route pattern (Meaning ICTs, SIP trunks, or GWs).

We do not have any way on which we can send a call back from CUCM unless the other end sends it back in or if a H323/SIP GW to hairpin it back to CUCM from there via dial-peers.

If you add a CUBE in the middle you could route it back to CUCM, but just a SIP trunk from CUCM to your PBX, no way.



if this helps, please rate

Phil Bradley Sun, 04/22/2012 - 07:10

I assume this is why the call redirect direct function is UCCX had the same issue when trying to redirect a caller to the SIP trunk. It would always hit the successful branch even if the SIP trunk port was shutdown.

I have noticed that CUCM will send a busy signal back to the caller after the set number of SIP Invite retries has timed out so I really don't understand why they do not use this in their logic to test for a trunk being down. I understand that the other end is responsible for other messages like busy or any other failed attempts once it reaches the SIP trunk but CUCM in this case knows that it cannot reach the SIP trunk after the invites time out. This is where you could then redirect the caller to some other destination.

As far as the CUBE goes I assume it has the ability to detect a SIP trunk down and then what? I would need it to route to a CTI route point in UCCX so it would go to a live operator. This seems like a bit overkill for the scenario above but if it's the only option then I will need to look into it. I guess you would need a CUBE also if you had redundant SIP trunks since call manager can't handle these functions? What does the destination under trunk in call manager do when you have multiple addresses? Does this just do load balancing and no failover? I wonder if the invites timeout will it send it to the second address or is this just used for load balancing?



Phil Bradley Tue, 04/24/2012 - 12:38

Just an update for those that may run into a similar scenario. We did find SIP pings under the SIP profile in call manager. We have version 8.6 so I am not sure if this was added or has been around for some time. The good news is that the call handlers in unity connection will receive information from cm based on these ping results. We also found that we can then route to an alternate number such as a cti route point in uccx.

Basically callers in our scenario go to an automated system under normal conditions that is via the SIP trunk. If for some reason the server is unavailable, not reachable via ping, we then route them to uccx for a live agent.


This Discussion

Related Content