First of all, I realized my load testing was actually on an AWS transcoder...
But this can be replicated in the devenv. Prereq is that we fix the current protocol deployment script so the protocol is in a running state after the deployment. (Currently it's paused by default, but @yondon is working on a new branch, hopefully it can be merged back in soon)
With that fixed, we can do the following:
- Deploy a fresh protocol
- Run a broadcasting node (also the bootnode)
- Run a transcoding node
- Stream a video into the broadcasting node (rtmp://localhost:1935). At this point, the transcoder should pick up the work.
- Run another broadcasting node that's connected to the network (Let's say it's running on
rtmp://localhost:1937) - Use
ffmpeg -i rtmp://localhost:1935 -vcodec libx264 -acodec aac -bitrate 1000k -f flv rtmp://localhost:1937/to replicate the original RTMP stream into the new broadcasting node. This should create a new job, and the transcoder should pick that up. - We can repeat steps 5 and 6 to add more concurrent jobs for the transcoder.