Drop nfs disk stats from telegraf monitoring on compute nodes

We don't need to collect the same set of ~1.9k stats on ~222 nodes every 60 seconds. This config will instruct telegraf to ignore nfsv4 mounts and improve the run-time of the metric collection step. Will remove ~41k log entries from syslog.

added 9 commits

• 445ae476...48c95190 - 8 commits from branch master
• 3ab976ec - Merge branch 'master' into 'fix-telegraf-disk'

Compare with previous version

Thanks @kerriw ... Looks good to me

merged

mentioned in commit 86b3ae9d