I looked at the monitoring results. Don’t see anything in the logs.
The only thing i can see was a CPU spike around the time the meeting started and continued until it ended. However it only registered 4 participants which is very odd.
The virtual service has 8 vCPUs assigned to try to mitigate this kind of behavior.
In contrast to your experience, - Thursday we had a 11 person meeting 7 video sources, 10 audio and CPU load only hit ~ 3.0 load. (We did have trouble sharing GOOGLE EARTH PRO as a screen share)
Observed from the described incident above.
CPU Load according to Linux did not exceed 6.0. Still high but not critical
: Hypervisor showed ~80-90% cpu usage on guest
Physical machine CPU usage did spike but stayed under 80%
Guest disk spike in usage ~32KBps but latency was under 8ms max
Host disk spike at the time ~500KBps but latency was under 8ms max
Bandwidth was 150/425 KBps (in/out)
RAM was at 10% of 16GB
Honestly not sure what could have caused this.
Occurrence 05:55 JST - 6:10 JST = 4:55 p.m EST - 5:05 EST
Host used iPhone to start meeting
Video continued to work
Monitoring shows 4 people (Need to confirm)
Echo Test completed fine
Audio exists for the first few mins before cutting out
Possible issue suggested
The overall CPU usage should stay below 70%. If the overall CPU usage consistently stays above 70%, you’ll likely get audio troubles as the BigBlueButton server will not have enough available CPU to keep up with the audio packets.
Analysis - I can only speculate that something happens on the guest machine to cause the CPU to spike affecting the audio. Currently no evidence of what that was.