Inconsistency in brick status reporting: stopped vs. down vs. down/unavailable

[julienlim]

This issue impacts multiple repos besides the UI repo.

Currently, when a brick is Stopped or it's unavailable, the brick is shown as Stopped in the Tendrl UI.

In all of the Tendrl/grafana Dashboards, the brick is shown as Stopped and Down.

"gstatus" reports the brick is as DOWN or down/unavailable.

gluster volume status doesn't show the brick that is stopped/down/unavailable at all, but you can deduce it from looking at the gluster volume list.

Tendrl UI screenshot - brick list and alerts:

Volumes dashboard (but same applies to the other dashboards which also contain the Bricks panel)

gstatus and gluster volume CLI output:

Need to resolve this before we can resolve https://github.com/Tendrl/ui/issues/899.

From a gluster perspective, there are currently only 2 states, i.e. it's running or it's not. There's no way to distinguish stopped vs. down vs. hung.

@gnehapk @a2batic @r0h4n @shtripat @shirshendu @cloudbehl @Tendrl/qe @mcarrano @jjkabrown1

[julienlim]

Related BZ that I just opened: https://bugzilla.redhat.com/show_bug.cgi?id=1566161

[julienlim]

After thinking about this some more, my recommendation is that for the 2 states for the brick, we should follow how gstatus does it, i.e. go with Up and Down to make it easier to understand for the user. See gstatus sample output examples.

Saying a brick is "stopped" when the brick is stopped, truly down due to failure or because the node is down, or the brick process is hung, etc. is not clear enough to the user that perhaps someone needs to look into the matter. Hence, why I suggest we go with Down or Unavailable like how gstatus reports it.

Thoughts?