Hi, I needed some guide on how can we autoscale jitsi on AWS servers
The theory for this is to identify the different scalable elements and their relation to each other as well to the entire system (jicofo, jibri, videobridge etc). Then you need to ascertains the metrics on which you want to scale (cpu load, total clients connected, total recorders etc etc).
In theory you would want to grab statistics from different components centrally and take scaling decisions accordingly. That is what an autoscaler does. Now aws/gcp etc offer this but they are only based on load percentage mainly and do not consider other metrics. which is why you should be looking towards ready to use solutions out there or github projects that you can reuse. in terms of jitsi, you will have component scaling then you will also have shard scaling (which can be quite a challenge). See the following references for help: