Today, I’m going to give you a troubleshooting tip about an issue I’ve been facing, on Cisco Finesse.
Agents started complaining that they suddenly get disconnected, and when you see the reports on CUIC, the reason is Connection Failure.
For this case, we are using Cisco UCCX 11.6.2.
First of all, we have to check the Layer 1. Make sure the phone is not losing connection due a cabling faulty.
If you are using Jabber, make sure you network connection is stable, and if it’s VPN, your internet is stable.
Voice traffic is really sensitive, so any minimum interruption can cause a disconnection.
Another thing Cisco recommends is, if your agent has Deskphone and Jabber configured with the same line (but not using at the same time, as UCCX does not support shared lines), you have to keep only one added to the End user and to the Application user. If you have both, it work, but you will have that disconnection some times as well (yes, I faced that in the past).
Now, the latest I’ve heard from them!
As per this Troubleshooting, these presence driven logouts occur when UCCX does not receive presence available status from the agent PC/browser. The system logs the agent out after 60 seconds.
So, seen all this points, there are 2 more difficult things to be caught, and I recently came across.
Chrome v88+ and Edge are known to cause these issues.
For agents logged out with the tab minimized/backgrounded:Disable Automatic Tab Discarding:
For versions 75 and above: Add chrome extension ‘Disable automatic tab discarding’ – https://chrome.google.com/webstore/detail/disable-automatic-tab-dis/dnhngfnfolbmhgealdpolmhimnoliiok
- Network LatencyOne of the Finesse requirements is the that the Network Latency cannot be higher than 400ms.
And that was exactly the problem I found on my network!!!
But how do we find out that the latency is going over 400ms??
Here are the instructions to gather the clientlogs from the agent Desktop side,
* Clear browser cache
* Load the following URL: <protocol>://<ip/host>:<port>/desktop/locallog and select “Sign In With Persistent Logging“. You will be redirected to login page with the appropriate query parameter url.
* Sign into Finesse
* Operate Finesse as usual
* When you run into the problem open a new window or a tab and reopen with same browser type using the following URL: <protocol>://<ip/host>:<port>/desktop/locallog and select Refresh button
* Now you have all the logs in the contents of the console output.
After analysing the logs, I could find the following:
Line 384: 2021-03-29T09:28:50.812 +02:00: 39DED1: <a href="http://<http://<<uccx_server>>: Mar 29 2021 09:28:50.728 +0200: Header : Client: 2021-03-29T07:28:50.518Z, Server: 2021-03-29T07:28:50.434Z, Drift: -84ms, Network Latency (round trip): 587ms
In this Log’s pieces, we can see that the roundtrip latency for the agent that was logged out, spikes above the 400ms threshold allowed by Finesse. This latency means that the server does not receive the “Presence available” notifications from the agent PC. After 60 seconds without receiving a notification, the system will log the agent out per design.
IntensiveWakeUpThrottlingEnabled Starting with Chrome 88: Improved resource consumption for background tabs To save on CPU load and prolong battery life, Chrome will limit the power consumption of background tabs. Specifically, Chrome will allow the timers in the background tabs to only run once per minute. If agents are using Chrome v88+, navigate to “chrome://flags” in the agent Chrome browser, search the above flag and ensure it is disabled (default=enabled).
So now you have to troubleshoot your network to find the source of that Latency.
That’s it guys!
I hope this post can help you out!