While the Monroe-Neumann criterion predicted dendrite-free all-solid-state lithium batteries (ASSLMBs) benefiting from the inhibition effect of solid-solid boundary and the high stiffness of a solid electrolyte (SEs). After the invention of inorganic SEs with high Li-ion conductivity, the experiments show that lithium dendrites can still penetrate SEs associated with SE cracking. This problem becomes counterintuitive, especially after the efforts in improving interfacial contact and the compactness of a SE (e.g., using a single-crystal SE) cannot solve it. Hence, the dendrite nucleation and penetration is likely an inherent problem associated with interfacial defects and the fracture resistance of a SE during ion transport. In this work, we developed a theoretical model to reproduce the recent findings in the in-situ experiments of ASSLMB failure during electroplating. The model is based on the framework of the phase-field approach with particular efforts to model interfacial defects (caused by electro-stripping), cracks (caused by internal stresses during ion transport), and lithium penetration (caused by an inhomogeneous electric field due to the existence of voids and cracks). With the model, we reveal the critical current density (CCD) and how it depends on materials parameters. We also explained the role of stack pressure and why it is ineffective in many experiments. Finally, we provide a stability map based on the possible toughness of SEs and residual compressive stress in them. The latter is, in our view, the route to solve the dendrite problem of ASSLMBs.