OSMS (should be similar to identify I'd imagine), 1 mm PTV, max iso to target distance 6 cm (we have a paper on this in revision). If necessary we split into two isocenters.
What motion is physics worried about? You can do the math and find that for short iso to target distances you need a lot of rotation for geometric miss. That much rotation should be picked up by surface monitoring. When that happens we assess why the patient is moving and repeat come beam, adjust 6D couch, and then resume treatment.
We picked 6 cm isocenter distance because once you get that far out you can get small rotational errors that are hard to detect that can move the mets more than 1 mm. You can do CBCT after or during treatment to confirm surface monitoring is reasonably accurate and math it out if you want to confirm met position vs. rotational changes as measured by CBCT.