Last week I was monitoring a relatively large vSphere cluster with several host clusters and datastore clusters. I saw a couple of datastore warnings related to the datastore usage. There were two or three datastores that were more than 75% utilized.
Strange, because other datastores in the same datastore cluster had more than plenty of free space. So why didn’t SDRS kick in to distribute the files more evenly?
The reason for this is that the default warning threshold for the alarm is 75%. However, the default space utilization ratio threshold for SDRS is 80%.
So SDRS only kicks in when the utilization is more than 80%. But by then, the alarm for the datastore usage is already triggered.
My solution for this ‘problem’ was to align both thresholds at 75%. If for some reason SDRS cannot move files to another datastore and it gets even fuller there is still the alarm alert that is set at 85%.