Now that I'm learning (and loving) the Admin API, I'm confronting an issue I've never felt has been particularly well addressed, at least not in any official capacity- the database cache. I've seen all sorts or recommendations on the size as a fraction of the available RAM, but I'm wondering if there is any new info on this. In particular I'd love to hear any updated suggestions for the cache size on 8GB and 16GB Windows Server 2016 deployments.
Actually perhaps I should start with questioning my core assumption, in case I'm way off base. The cache is the amount reserved for FMS, but not the max it can use right? Is it there "just in case" the server (let's assume a dedicated FMS deployment) starts gobbling up RAM? Why not just leave it at the default and let FMS use what it needs?
Based on input from the community and consultants I've used, I had been specifying 3GB on 8GB deployments.