Stalling tolerance

[tmigot]

Is it possible in Knitro to fine-tune tolerance leading to the following conclusion:

Primal feasible solution estimate cannot be improved; desired accuracy in dual feasibility could not be achieved. No further progress can be made. It’s possible the approximate feasible solution is optimal, but perhaps the stopping tests cannot be satisfied because of degeneracy, ill-conditioning or bad scaling.

See the doc