Hey everyone! I’m the network engineer and I’ve been tracking our voice quality for remote agents. I’ve noticed a strange correlation. Our agents’ MOS scores are consistently dropping from a healthy 4.2 to a struggling 3.5 whenever ‘Screen Recording’ is active for the interaction. We are using the standard Genesys Cloud desktop app and the agents are on home fiber connections. It looks like the screen recording upload is competing for bandwidth with the WebRTC media stream. Has anyone else seen this performance degradation, and is there a way to prioritize the voice traffic over the screen recording upload within the desktop app itself?
Hello. I build flows for many different clients and I have seen this bandwidth competition multiple times, especially in regions with poor outbound internet speeds. The Genesys Cloud desktop app utilizes the same local network stack for both voice and video. To fix this, you must implement QoS (Quality of Service) tagging on the agents’ machines. You should tag the voice traffic with DSCP 46 (EF) and the screen recording traffic with a lower priority like DSCP 0. This ensures that even if the upload pipe is saturated, the voice packets are always prioritized by the local router. It is a mandatory architect-level setting for any large-scale remote workforce!
I maintain over fifty flows and I’ve heard this complaint from our supervisors constantly! They think the platform is ‘broken’ but it’s just a network bottleneck! In addition to the QoS tagging, you should also check the ‘Screen Recording Resolution’ in the recording policy. By default, it might be set to a high bitrate. Reducing the resolution to 720p at 5fps can significantly reduce the outbound bandwidth requirement without making the recording unreadable for the QM team. It’s a much easier ‘quick fix’ than trying to coordinate QoS settings on a thousand different home routers!