Claudio Costa

That's good to know. I haven't looked in this yet but I think we'd need changes on this side as well because at the very least we need to understand...

I think we are still stuck at https://github.com/electron/electron/issues/37524 although I haven't really had a chance to test on Wayland still to see if there could be workarounds. Our internal ticket...

FWIW I am hitting this same issue when trying to receive multiple differently encoded tracks.

@Sean-Der Any thoughts on this? We may actually need to make similar changes on our end so it would be great to get some feedback on what could be the...

> Hey @streamer45 > > Is this what you are trying to implement? You want to have one `TrackLocal` that can be used to send many different codecs? Since you...

Weird, I would expect to see it in there unless `GOTRACEBACK` is being used. May be worth checking the service file. @PalachN what operating system version is Mattermost running on?

@PalachN Thanks. To debug this further it would be helpful to share the following: - The Mattermost `config.json` file (sanitized). - Your `mattermost.service` systemd file.

I did some quick bisecting and it seems that this has stopped working in Firefox 96 so something must have changed [on that side](https://bugzilla.mozilla.org/buglist.cgi?f1=cf_status_firefox96&o1=equals&query_format=advanced&v1=fixed&order=bug_status%2Cpriority%2Cassigned_to%2Cbug_id&limit=0&component=WebRTC&component=WebRTC%3A%20Audio%2FVideo&component=WebRTC%3A%20Networking&component=WebRTC%3A%20Signaling), just not sure what exactly.

> Ok, then we are deviating from what `h.connectionCount` is tracking. `h.connectionCount` is tracking all "active" webconns, which means only registered ones which currently have a connected client. But `mi.HTTPWebsocketsGauge`...

@agnivade I updated the release note and also planning to review our documentation prior to the release to reflect the slight change in semantics. Let me know if there's anything...