STOR2RRD monitors IO and data throughput of all volumes and creates graphs per volume and aggregated with all volumes. In order to make aggregated graphs easy readable it comes with its own approach how to place all volumes into single aggregated graph.
There are several approaches for volumes name conventions. Two most common approaches with impact on presenting them by STOR2RRD will be explained here.
- All volumes assigned to one server have same name (alias)
STOR2RRD presents only names and groups volumes which share the same name. However you can get each volume statistics if you want via links which are on the detail page of each volume name
Volume names are unique per storage.
STOR2RRD here presents each volume separately in any graph
Volume aggregated graphs
There are graphs where all volumes are aggregated. They are data rate and IO rate.
As you might guess to have thousands volumes in the one graph does not sound like a good idea.
Therefore STOR2RRD implements filtering which significantly reduce number of presented volumes in one graph.
Filtering is based on IO or data limit. There is general IO or data limit which defines which volumes will be presented in the graph. Rest of volumes which do not reach that limit are grouped and presented under one colour.
There are 2 kind of limits:
- Global one in etc/stor2rrd.cfg
VOLUME_IO_MAX=100 # in IOPS
VOLUME_DATA_MAX=2048 # in Kbytes
- Limits per storage in etc/storage-list.cfg
Note that limits are about peak IO or data throughput.
# data/io limits for displaing volumes in aggregated graph
# defauls are set in etc/stor2rrd.cfg: VOLUME_IO_MAX=100; VOLUME_DATA_MAX=1024
Where there is no limit set then the global one is taken.
As above you can see that general limits for storages is 2MB and 100 IOPS but storage DS05 has its own limit set to 1MB and 50 IOPS.