The thing is, when I start a transmission with 7 streams on the encoder, it arbitrarily delays the ingestion between the streams of that transmission (assuming that one stream is a bitrate quality). This delay is a normal behavior since encoders have a long bootstrap for every bitrate.
Said this, if i just add one stream after other like Bryan suggested, it will probably diverge their media-sequences. To fix this, I destroy the entire stream group and create it again with all streams of that group (createHLSStream passing all streams on localStreamNames) to achieve media-sequence synchrony. It’s working this way, but we’re noticing cases where one bitrate is missing on the master playlist.
We could identify the problem: Imagine that the first stream has entered and after 1 second another 3 enters at the same time. On the time that this 3 streams are ingested, Evohandler needs to destroy+create for every single bitrate introducing a race condition problem (since destroy+create is not a atomic action).
If we could just reset the media-sequence of a stream group, this problem would be solved.