We had a videoconferencing 'media node' that we tested with our Bot Army. Overnight (or all day or all week) we had 100 headless clients clicking into and out of conferences at random. Turning mic and video on and off.
First it ran for 10 minutes and crashed the media node. When it reached an hour we celebrated! Got it to stay up for days eventually.
Which meant, what? No more bugs? Well, it meant the average user would not encounter them in practice with maybe years of use. Good enough.
We had a videoconferencing 'media node' that we tested with our Bot Army. Overnight (or all day or all week) we had 100 headless clients clicking into and out of conferences at random. Turning mic and video on and off.
First it ran for 10 minutes and crashed the media node. When it reached an hour we celebrated! Got it to stay up for days eventually.
Which meant, what? No more bugs? Well, it meant the average user would not encounter them in practice with maybe years of use. Good enough.