30 ohms is a lot to ask from copper, at normal design rules. 100fF is fine.
The tempcos of both are rather extreme: copper is, what is it, a bit worse than PTAT I think? And the capacitance is whatever the laminate is, so depends on whether you're doing sandwich (traces above/below) or adjacent (traces on same layer), and obviously there's no shielding to surroundings.
The other reason Cu sucks for resistance is its inductivity. Typically reactance dominates above 10-100kHz, even for structures that attempt to cancel out inductance (e.g. hairpin loops instead of spirals). So it's hopeless as a current-sense resistor (where the small resistances might prove useful) in switching circuits.
Or did you mean embedded components as in carbon ink, chips embedded in the stackup, etc.? Which are also possible -- but I can't imagine a proto fab doing such esoterica..? I mean unless that's what they've said they can do, I haven't checked anything.
Tim