
I will discuss three models containing relativistic hydrodynamics: N=4 SYM at strong coupling, Mueller-Israel-Stewart theory of second order hydrodynamics and its new generalization. I will demonstrate, using Bjorken's boost-invariant flow, that in all three models the hydrodynamic gradient expansion is divergent due to the presence of exponentially decaying modes not captured by the hydrodynamics gradient expansion. I will then discuss how to properly think about the "resummed hydrodynamics" and show that in expanding plasma systems described by Mueller-Israel-Stewart theory such resummation does not produce significant deviations from the gradient expansion truncated at the first (viscous) order.