I can't tell whether you've even addressed my concern, honestly.

We can engage in a process of reductive analysis to find a seemingly most-primitive set of things-that-exist. Yay! We have some beables now. No arguments there.

However, what I question is the move to declare that relationships have any dependence on observers; and further that the magic pixie dust we call "ontology" can be sprinkled upon our beables but not upon properties arising from relationships.

(At a QM level, relationships are beables, too, at least in the DeBroglie/Bohm formulation. We needn't worry about that here, however.)

This-happens-because-of-that-relationship is no different in kind that this-happens-because-of-that-physical-property. Indeed, almost everything involves both: this has charge, which makes an electric field, which induces a force on that thing which is charged.

What I don't understand is elevating one half of the equation to exalted status as "ontology" and neglecting the other to the point of doubting that it counts if consciousness is implemented in the other.

The "observer" thing is dispensable, you know. (See above link.) The math doesn't demand an observer. There are at least two ways to get out of itt. You can use the DeBroglie/Bohm formulation and the observer is gone. Or you can declare an "observer" to be just any old macroscopic assembly that an observable gets entangled with, which has exactly zilch to do with consciousness or intersubjectivity.

Fundamentally, very many relationships have functional consequences.

At some point, one can define ontology such that local beables are ontologically novel and other things aren't, but eventually one has to step back from that particular language game and ask: who cares? How does that matter at all?

It seems like the entire issue is granting infinite computational power to infer every consequence on the one hand, then taking it away with the other, noting the disparity and going, "welp, consciousness can't be explained like that".