Hi!
We have a SaaS where caregivers meet patients, and we offer the video session functionality as part of the subscription, in web and app.
We have openvidu on premise, one master and one node (v.2.28) (we will increase nodes when we have users)
I had my first longer video session with a potential customer today to demo the service. It was in the desktop browser, I was using Chrome on Windows and the other had Windows Edge.
I could hear and see the customer the entire session - no issues.
They reported that my audio was hacky 3 times and my video got stuck 2 or 3 times. It sorted itself out my itself.
We have screenshare functionally so I had 2 openvidu sessions and so did the customer (but I had issues with hacky audio before I started to screenshare.
The master CPU is almost steady at 25% and didn’t increase during our session.
The node CPU is usually at 5% and increased to 16% during our session.
The openvidu inspector analysed the session and it reported no issues at all.
I’m thinking that the customer might have had issues with Edge or the computer.
I need to test a longer call with a colleague or friend before I do it with a customer again.
And so to the question: If I experience the same thing again, its laggy and the servers “look fine”, how do I best troubleshoot this? With your experience, what’s the first action you would take? How do I point to their own setup? 
We are both based in Sweden and the servers are based here as well, inbetween us.
Thanks in advance,
Richard