The mechanics of the solver for different governing equations are quite different, so it isn’t appropriate to compare the robustness of changing the time step in a quasistatic linear elastic simulation to quasistatic poroelasticity.
The time scale used for nondimensionalization (normalizer.relaxation_time) is important for poroelasticity. As mentioned in PETSC ERROR when use simulations in Pylith 4.1.3 - #6 by baagaard, for a well-conditioned system for poroelasticity, the ratio of the nondimensional permeability to viscosity should be about 1.
Another potential issue is that the default time stepping algorithm for quasistatic simulations is backward Euler. A higher order time stepping method may work better. This is something we have not yet played around with.