[usage] Add usageAttributionID to WorkspaceInstance model (in go)

[easyCZ]

Description

Adds model definition to consume usage attribution id

Related Issue(s)

• Part of Use attributionId when listing workspace instances for usage #10922

How to test

Unit tests

Release Notes

NONE

Werft options:

• /werft with-preview

[jankeromnes]

Many thanks for this PR!

In general, I'm very much in favor of multiple smaller PRs. However (and maybe this is just me adjusting to this new paradigm), I still struggle with reviewing code that isn't used in any meaningful way.

Thus, I took a peek at your other PR: #10938 👀

• There, I saw how the attribution.Values() function is being used 👍 (helps me build confidence that it's indeed what we want)
• However, I also saw there that the NewAttributionID(entity string, identifier string) is not the most useful method, proven by the fact that you've refactored it into db.NewUserAttributionID(identifier string) and db.NewTeamAttributionID(identifier string) -- would it make sense to hoist the better implementation in this PR?

[easyCZ]

I still struggle with reviewing code that isn't used in any meaningful way.

Here, the constructor and the destructor are mostly for completness and symmetry. Agree that ideally the destructor wasn't used in this PR (and perhaps should've been used in tests) but it felt incomplete without it.

However, I also saw there that the NewAttributionID(entity string, identifier string) is not the most useful method, proven by the fact that you've refactored it into db.NewUserAttributionID(identifier string) and db.NewTeamAttributionID(identifier string) -- would it make sense to hoist the better implementation in this PR?

This would also suffer from the same problem as above. I'd be creating a constructor for NewTeamAttributionID without having usage for it in this PR, so that felt even more of an over-kill to include in this one.
On the other hand, once I had practice usage, it longer made sense to keep the NewAttributionID function public so it got refectored. This mostly shows that you raise PRs as you go, rather than figuring out what the perfect final version is (and back-slicing it then).

[easyCZ]

I've pulled in the two methods and added some tests for using the Value to showcase it better. Hopefully this addresses the above comment.

[jankeromnes]

Looks good to me, many thanks! 🚀