Disk graph related information are taken from “df” and “cat /proc/diskstats” outputs.
In particular, “Disk Space” graph is created from “df” output by a filesystem provided.
All other disk graphs are created from the information from /proc/diskstats. This file contains per-block-device data.
So there is no easy way to get volume mapping and combine device data.
Theoretically, it’s possible but a volume mapping is not always available.
The following might be useful: if you need to graph a raid volume, say md0 consisting of sda1 and sda2, you can create “disk space” graph for “/dev/md0” and graph any of raid’s devices with all other disk graphs, e.g. graph just sda1 as sda2 will show exactly the same. The only thing you should consider it’s 1 of N equal devices of raid volume.