There has been a lot of research on time-varying failure rates, which deems constant failure rates as inadequate to model failures accurately. However, besides time, failure rates can also be affected by the state of the system (or its history, in terms of sequences of states and events that it has been through). In our paper we define several classes of state-varying failure rates and extend the formalism of Petri nets to model them. We further use the flexibility of the proxel-based method to accurately analyze behavior of systems that incorporate these kinds of failures. To illustrate our approach and study the effect of such dependencies, we compare simulation results for two models: one that exhibits state-varying failure rates, and another that only contains predefined failure rate functions.