Add component for empty MiniTable

[charliepark]

This adds an empty state for MiniTables. One hope is that this resolves (or at least mitigates) #2627.

The no-whitespace-diffs view is a bit easier to glance at: https://github.com/oxidecomputer/console/pull/2811/files?diff=split&w=1

Closes #2661

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
console	✅ Ready (Inspect)	Visit Preview	Apr 29, 2025 7:18pm

[charliepark]

Once we're happy with the approach, I'll roll it out to other MiniTable instances.

[benjaminleonard]

I know @david-crespo likes components to be more declarative, but I wonder if the current mini table implementation leaves room for accidental differences to sneak in.

I say that here because we could pass items to the mini table and let it handle the empty state within the component automatically.

We're also going to have two versions with these webhook changes, one that is pretty much the same as it is now and one with integrated inputs. In this case would we keep the existing component and create a wrapper component that takes props directly instead?

[charliepark]

I hear you on the concern about accidental differences, though I think the flexibility is also useful. For example, in the webhook subscriptions, we have this design:

The body copy there uses a different resource name (event subscription) than the title (subscription), and overall that block uses different copy than what's currently in this PR (No disks / Add a disk to see it here).

We could figure out unified copy for the empty table, but I think it might be good to leave it flexible at first and then see if there are dominant patterns we should reinforce with templated copy. As always, open to thoughts on that.

I'm not as familiar with the new dynamic webhooks patterns and what we'd pass to a wrapper component, but I suspect it'd be pretty straightforward.