Add to sizing minimum requirements info already in cbbackupmgr docs

At the end of the https://docs.couchbase.com/server/current/backup-restore/cbbackupmgr.html#discussion section, there's info on the minimum and recommended hardware for running cbbackupmgr CLI.

For cbbackupmgr CLI:

Ask is to expand on the existing guidance a bit more to explain using --threads (if the machine number of cores is 16, try setting --threads to 16, etc.), about monitoring memory, cpu, disk, and to improve throughput, increases resources (if possible) for whichever component is the bottle neck (memory, cpu, disk, etc).   

For the Backup Service:

Similar info as for the cbbackupmgr CLI – minimum and recommended hardware, that if the cpu, memory resources are increased, that the Backup Service will automatically adjust to use the additional resources available (when the next backup task/job runs).

For direct backup to S3 (applies to both cbbackupmgr CLI and the Backup Service):

There is some nice information in the Connect 2020 talk on Direct Backup to S3 – https://www.youtube.com/watch?v=VNNfObxkJeM  (talk starts at time mark 6:44) – consider if some of the info presented can be added to docs section on cloud object store backups – especially to explain a bit about the network I/O factor in performance.  The performance chart description starts at time mark 11:05 in the video – the chart is really nice since it answers the question about "how does this direct backup to S3 compare to just backing up to disk and then copying the backup files to S3" – and the chart shows data for different machine sizes and number of --threads used (for cbbackupmgr CLI).