[Observability] [SLO] Overview stats need better clarification and distinction between states

[benakansara]

On SLO overview page, when an SLO is both stale and one of healthy, violated, no data or degrading, the overview stats don't reflect it correctly. The stats don't include SLOs that are stale which is misleading as the UI shows healthy/violated/no data/degrading SLOs but stats says they are 0.

Acceptance criteria

• Healthy, violated, no data and degraded stats should include stale SLOs as well if SLO is both stale and one of healthy, violated, no data or degraded

[elasticmachine]

Pinging @elastic/obs-ux-management-team (Team:obs-ux-management)

[mgiota]

@benakansara I was thinking about this. If we want healthy, violated, no data and degraded stats to include stale SLOs, we need to be in sync with the SLO list as well. But doesn't this come against what we currently have, which is hide stale SLOs by default?

I was thinking that we could keep the current behavior and hide stale SLOs by default, but update the stats, when user clicks on the Stale stats count. In that case user sees the stale SLOs in the SLO list, so only in that case we can update the stats to show which of these stale are violated, healthy and so on. What do you think? Would this work?

As discussed the confusing part is that all stats represent the non stale SLOs and then we have a Stale SLOs number. These could be separated, for example show the non stale SLOs in one box and stale SLOs in another box. In each box then we could show how many of them are healthy, violated etc.

[jasonrhodes]

A couple thoughts here talking with @drewpost:

• The back button doesn't work when you filter by these statuses, we should fix (push not replace if possible)
• It looks like we already hide "stale" SLOs in this view, but show the count of them, and then when you click to filter by "stale" they appear?
• These statuses are NOT distinct, is that right? If that's true (meaning, an SLO can be BOTH "stale" AND "healthy", etc.) then should we reconsider displaying these as siblings in the view shown in this issue description? That seems confusing.
• Once we answer these, can we revisit what we are trying to do with this specific issue?

[jasonrhodes]

As this hasn't been refined, I'm removing it from the iterations. @drewpost can you sync with engineers to either get this refined and ready or moved to backlog if we aren't moving forward?