The observation of the first pulse-to-pulse glitch in the Vela pulsar opens a new window on theoretical speculations about the internal dynamics of neutron stars, as it allows to test models for the description of the first moments in a glitch. We improve the analytical study of the minimal three-component model for pulsar glitches, by solving it with generic initial conditions for the two initial lags of the superfluid components. The purpose is to use this solution to fit the data of the 2016 Vela glitch - by employing a Bayesian approach - and to obtain a probability distribution for the physical parameters of the model and for observational parameters like the glitch rise time and the relaxation timescale. A physically reasonable, non-informative prior has been set on the different parameters of the model, so that the posterior distribution can be compared with state-of-the-art information obtained from microphysical calculations. By considering a model with a tightened prior on the moment of inertia fractions and by comparing it with the original model by means of Bayesian model selection, we study the possibility of a crust-limited superfluid reservoir. The posterior distribution for the moment of inertia fractions of the superfluid components, the coupling parameters and the initial velocity lags between the components has been obtained. Analysis of the inferred posterior also confirms the presence of an overshoot in that glitch, and sets an upper limit of 6 seconds on the glitch rise timescale. The comparison between the two models with different priors on the moment of inertia fractions seems to indicate the necessity of a core participation in the glitch phenomenon, regardless of the uncertain strength of the entrainment coupling.