Skype for Business is new to many organizations, and for many IT leaders deploying Microsoft's unified communications tool, it will be their first time working with it. When we are working with organizations to optimize their Skype for Business deployments, we regularly see the same questions around troubleshooting. Issues will arise in Skype for Business, as they do in other vendor platforms, such as Cisco and Avaya, but a typical issue in Skype for Business can be different because the nature of the platform. Here I'm going to take you through three areas you should start your troubleshooting with: this is Skype for Business Troubleshooting 101.
1. Checking Server Health when Troubleshooting Skype for Business
The first area you should start troubleshooting is your server health. Poor server health or memory saturation of server clusters will negatively impact Skype for Business performance.
Start by looking at your server health from a high level and then drill down to those in the red and see what areas are being affected: latency, jitter, etc. This will give you a better understanding in terms of resource saturation and thresholds.
The questions you want answered with troubleshooting your server health is: what's eating up your resource? Is it a specific process? How many CPUs have been used up? How much physical memory and virtual memory is being consumed? Companies with virtual memory will see when the memory is hitting peak saturation levels allowing them to allocate additional resources accordingly.
2. Conference Call History
"What was wrong with the conference call?" A familiar question to those in charge of troubleshooting unified communications and one which can be answered through checking the conference call history. From your home dashboard drill into completed calls and conference call history. In Prognosis you can see how all the details are stitched together and displayed in a tabular format. Select your call and identify the number of participants on the call. Straightaway you can determine whether it was one user having issues on the call or multiple users i.e. a broader reaching issue.
In the example shown below you can see of the three participants, but only one person had issues: Oliver has multiple threads to the front end whereas Menage and Michelle don't.
In IR, we saw this play out in real life when an IR customer deployed Skype for Business. A large conference call took place and everyone was having difficulty understanding what the presenter was saying. The assumption was that it was far reaching infrastructure issues as everyone was experiencing poor quality but instead it was the originator and presenter of the call, the VP, who was having poor performance due to a less than optimal mix of technologies. Like the attendee in the example, Oliver, the performance was poor - he was on and off and on and off, not anything else.
In Prognosis you have the ability to drill down into each stream and the kind of things we learn from what we are pulling back is (and what we can drill into is):
ID, endpoint, whether on wired of VPN, statistics on when the user joined and when it was terminated, and the type of server (an edge server) which was remote, the minimum MOS in each direction (which is really important, as you could have a great average MOS but as an end user you are going to experience the worst), and the details Microsoft give; latency, jitter, packet loss.
In the conference calls details the signalling session was terminated due to a loss in network. So essentially Oliver was off campus, using an unreliable network and that's why he was falling off and coming back on.
A good lesson to take from this example is not to underestimate the value in educating end users. If their mindset shifts from “the conference call doesn't work properly” to “I wonder if my network/server/endpoint is working…” you'll make large strides in your user adoption of Skype for Business, which is one of keys to success.
Lastly, if you're wondering what happens if some of the people on the call weren't using Skype for Business, but Avaya or Cisco instead: wonder no more! A third party multi-vendor monitoring and troubleshooting tool, like Prognosis, will be integrated with these platforms, so you can drill down into their instances as well.
3. Troubleshooting Skype for Business In-flight Conference Call
The final area where you might want to start troubleshooting Skype for Business calls is real-time “in-flight” conference calls. This is helpful if someone contacts the helpdesk while they're having problems on a conference call. Prognosis provides the ability to navigate towards the particular active conference call that's underway and drill down to see what's happening. Microsoft send information every 30-45 seconds so the dashboard is getting updated in real-time and the administrator can see whether the issue is isolated to the one user, or a broader issue, just like in historical calls. It can see the underpinning metrics of latency, jitter and packet loss and the overall call status. A color system can show whether performance is good, average or poor on each of the elements.