It would precede rupture. There is a loss of material ductility as the triaxial stress state increases, whereby the material behaves in a brittle manner and fractures 'locally'. This failure mode typically requires some form of strain concentration, and the original method was validated using notched bar and tensile tests. There doesn't need to be a purely triaxial stress state for strain accumulation, so it depends on the degree of triaxiality, which is why you need to follow the method and not use a single strain limit (particularly not the ultimate rupture strain).

Having said that, I believe a suitable screening strain limit could be developed, relevant to the PE applications where this method is employed, but that is another discussion.

There is not much validity to the elastic method, so I would avoid using this for anything more detailed.

Thank you BJI. Very crearly.