I have use case to scale pods (HPA) based on memory (and custom metrics).
However respective apiVersion is in beta (autoscaling/v2beta2).
Not sure when this will be in GA/stable version, hence checking.
Any advise for me?
If you look at the Kubernetes API Versioning doc and the API deprecation policy doc there is no clear indication as to when autoscaling/v2 will be available (GA).
For example, in the next Kubernetes release, the API could be either autoscaling/v2beta2 or autoscaling/v2 depending on how the release/feature team feels about it in terms of satisfying 'Stability'
Having said that based on the deprecation policy and rule 4a here, you should be safe to use autoscaling/v2beta2 (for at least the next 3 Kubernetes releases after autoscaling/v2 or autoscaling/v2beta3 is announced.
So let's hypothetically say autoscaling/v2 is released with K8s 1.19, then autoscaling/v2beta2 would be available in Kubernetes 1.20, 1.21 and 1.22.