energy
is the value of the Hamiltonian (up to an additive
constant) at each sample.
While divergences can identify light tails and incomplete exploration of the target distribution, the energy diagnostic can identify overly heavy tails that are also challenging for sampling. Informally, the energy diagnostic for HMC quantifies the heaviness of the tails of the posterior distribution. The energy diagostic plot shows overlaid histograms of the (centered) marginal energy distribution and the first-differenced distribution. Keep an eye out for discrepancies between these distributions.
For more details see Betancourt, M. (2017). A conceptual introduction to Hamiltonian Monte Carlo.