Disk usage grafana dashboard panel improvements

[Andrea Ieri]

The Telegraf host metrics dashboard contains a disk usage panel that could use a couple of improvements.

1. the Y axis is confusing

The PromQL query is `disk_used{host=~"$host"}/1000/1000/1000`, while the Y axis has no unit. We should let Grafana figure things out: set the query to simply be `disk_used{host=~"$host"}`, and configure the unit as bytes (via the GUI this is found in Field → Standard options → Unit. The panel title should of course also drop the unit.

2. too many series

Reporting on the snap disk usage is useless. We should filter out squashfs filesystems.

[Eric Chen (eric-chen)]

Soon or later, the monitoring system will migrate to COS. Telegraf will be replaced by grafana-agent. Therefore, we won't follow up this new feature request
(https://charmhub.io/topics/canonical-observability-stack)