Memory aware scheduling

Hi, I would like the autoscaler to utilize all the resources while using ActorPools. Essentially I create a pool of actors by specifying memory requirement per actor. However, the resources do not seem to be fully utilized as I can see free memory in the workers where more actors can be initialised.

I also want to profile/log the total resource (specifically memory) usage throughout the cluster lifecycle. Is there a way to access this information?

Thank you!