Courtesy Callback Reporting HA

[bankspb]

I have two side - SideA and SideB, two Call Servers, 2 Unified CVP Reporting Servers.

On Courtesy Callback Configuration page I can associate only one  Unified CVP Reporting server SRV-CREP11A or SRV-CREP11B. if SRV-CREP11A is down, CCB isn't work!

How setup automatic redundancy for Reporting Servers??? If sideA is down, I need for work CCB thought sideB

[geoff]

The requirement that the CVP Report Server be collocated with the Call Server / VXML server has been relaxed. Now you only need one, as long as the response time from the other side to the side where the Report Server lives is 80ms RTT. So this makes it a lot easier to report on the internals of Studio apps. Most deployments around 2000 agents are just going to have one.

Now onto your question - no, CCB is not highly available. If the Report Server is down, the web service probe issued by the Studio app fails, and callback is simply not offered. Perhaps not ideal, but not the end of the world.

If you want a Cadillac CB solution, buy Virtual Hold.

Regards,

Geoff

[bill.king1]

We had opened a defect on the colocation requirement a while back so that it would be updated in the documentation. We were also at the time that in addition to the round trip time requirement, there is also a maximum number of events that the one Reporting server can handle at a time (i.e. no more than X call events per second). I'll try and dig it up, but I'm not sure what the behavior is if you exceed that maximum number of events (i.e. it just discards the rest, you get alerts if it happens, etc.).

[geoff]

---

@bill.king1 wrote:

... in addition to the round trip time requirement, there is also a maximum number of events that the one Reporting server can handle at a time (i.e. no more than X call events per second)

---

Thanks Bill - I did not know that limit, so please post if you can find it.

Regards,

Geoff

[bill.king1]

The documentation defect still isn't viewable (not to me at least), but at the time we were told it was max events (not calls) per second had to be less than 420, it is touched on very briefly in the old sizing guides. In that same guide, they outline how there are some elements that are only a few events, but others as many as 30, so depending on your flow, 420 may be fine or it could be tight.

Let me know if you want the URL for one where we saw it and I'll post it. I'm just curious as to what happens if you exceed that number.