Session: 0t3dzlnmQD2+lvkys/EMmw==
VideoSDK for web version: 2.1.10 (latest)
I had some network stability issues in this meeting, such as frozen images and streams turning off. I was wondering what are the best practices to mitigate these issues, in such a way that we can have a similar experience as the Zoom app?
Hi @fides-it
Thanks for your feedback.
0t3dzlnmQD2+lvkys/EMmw==
After analyzing the logs, we found the iOS Safari user might be in a poor network environment with very limited bandwidth and a high packet loss rate, leading to video freezing. After refreshing, the situation improved.
We will continue optimizing the experience in weak network conditions in future versions.
Thanks
Vic
Hi Vic,
thanks for the swift response. The suggestions you made however are not sufficient for us. Customers are not happy and we need to do something about it. We can’t ask them to “just refresh” or “sit tight for future release of Zoom Video SDK”. That just doesn’t cut it. Unless you’re saying Zoom Video SDK for Web is not suitable for production use at the moment?
A few follow-up questions then if you don’t mind, maybe we can work it out.
Today there was another meeting with multiple problems for different users (EFBwgZXkQi2b8P1pMF9NBQ==). I spent some time analysing what’s happening:
User 1 always runs into trouble after 30-40 minutes into a meeting and today was no different. To me it looks like CPU related issues: spikes in “Device from web client”, but no significant spikes in MMR traffic. I suspect this user’s desktop (Windows/Chrome) is crashing with high CPU. This matches with the fact that I log all kinds of events to Sentry.io, but these events stopped during the “blackout window” (08:29-8:36 CET). Exactly at the time the freezing / video problems occur.
User 2 (MacOS / Safari) did not have audio input or output. Couldn’t hear anything, and we couldn’t hear her.
User 3 (MacOS/Chrome) ran into problems when she joined the meeting half way (around 08:20 CET). Was disconnected a few times. Then after 3 times, she was ok but in the end again one issue.
User 4 was disconnected at 09:55. Big spike in “Receiving - Device from web client” around that time.
To me it looks a lot like rendering issues on the device, perhaps due to CPU spikes. Would it help to set the video_webrtc_mode / audio_webrtc_mode to 1 to force the usage of WebRTC? Do you have other suggestions?
Also, why are the CPU stats empty in the session’s dashboard? It would help in this case.
Hi @fides-it
Thanks for sharing detailed info with us.
Would it help to set the video_webrtc_mode / audio_webrtc_mode to 1 to force the usage of WebRTC
Yes. WebRTC audio/video will improve the performance.
EFBwgZXkQi2b8P1pMF9NBQ==
We are working on the analysis of this session and will keep you updated.
Thanks
Vic
Hi Vic,
any update on this issue?
Best regards
Dennis
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.