I'm extending an existing ES cluster and am looking for general hardware recommendations e.g. of the form listed here.
That document is now a few years old and I assume some of the guidance has been updated, but I haven't found a more recent version. The closest I could find is a document on Elastic Cloud Enterprise, but I'm not sure how relevant its contents are.
In particular, I have the following questions:
- Is it ok to create an ES cluster from a combination of bare metal machines and containerized instances, or is the disparity in performance likely to cause issues?
- In general, should the hosts in a cluster be as similar as possible?
- Does the recommendation against placing indexes on a remotely mounted filesystem, e.g. NFS, still stand?
- Is there a recommendation against RAIDing local disks used for indexes?