Monitoring BTRFS and display select set of errors in a graph

[davurs]

Hi,

we're running Zabbix in a mixed environment, with one central server and a number of proxies. Some of our monitored Linux devices run of btrfs filesystem and from time to time we may experience errors in the filesystems.
I'm thus working on a way to display the number of certain sets of errors in a given BTRFS setup.

btrfs device stats /path/to/Datastore

displays a lot of info like write_io, read_io errs etc.
With some awk and sed, I've come so far on command line as to pick out one set of interest, let's say write_io errs - it yields something like

sda 26
sdb 145
sdc 120

Question: If I get this output from a UserParameter in zabbix.agentd.conf, how do I process it and put it in a graph on the Zabbix server?

Thanks in advance