I am testing version 2.15.
What exactly does ‘Load’ mean?
Numerically, when users joined, the media server was not added even though the load is higher than the preset ‘Max Average Load’.
Additionally, I wonder if there is a way to boot a media node that is off as the load gets closer to’Max Average Load’ in the’on premise environment’.
Load is measured as “CPU load” of the media node.
If you deploy OpenVidu PRO on premises, you can react to webhook events requesting a new node and informing an existing node is no longer needed and can be turned off.
You have more information here: https://docs.openvidu.io/en/2.15.0/openvidu-pro/scalability/#autoscaling
thanks…Micael.
So do you mean that in the ‘On premises’ environment, I have to manually check the ‘Load’ value of the active media node and add or delete another media nodes manually by my own API and calling webhook ?
No.
If you deploy OpenVidu PRO on premises and you want to use autoscaling, then you will receive in the webhook a notification when a new media node is needed (so you can add it to the cluster using the REST API) and when an existing node is not needed anymore (so you can remove it from the cluster). You don’t have to query any metric, OpenVidu PRO will do for you and notify your app when needed.
Regards
