The $S_8$ tension is a longstanding discrepancy between the cosmological and local determination of the amplitude of matter fluctuations, parameterized as $S_8equivsigma_8(Omega_m/0.3)^{0.5}$, where $sigma_8$ is the root mean square of matter fluctuations on a 8 $h^{-1}$Mpc scale, and $Omega_m$ is the total matter abundance. It was recently shown that dark matter (DM) decaying into a massless (dark radiation) and a massive (warm DM) species, with a lifetime $Gamma^{-1} simeq 55~ (varepsilon/0.007)^{1.4}$ Gyrs -- where $varepsilon$ represent the mass-energy fraction transferred to the massless component -- can resolve the tension. Thanks to a new, fast and accurate approximation scheme for the warm species, we perform a comprehensive study of this 2-body decaying DM scenario, discussing in details its dynamics and its impact on the CMB and linear matter power spectra. We then confront the robustness of the resolution to the $S_8$ tension against a number of changes in the analysis: different $S_8$ priors, marginalization over the lensing information in Planck data, trading Planck high$-ell$ polarization data for those from the SPTpol collaboration, and the inclusion of the recent results from the Xenon1T collaboration. We conclude that the preference for decaying DM, while entirely driven by the local $S_8$ measurements, does not sensibly degrade the fit to any of the cosmological data-sets considered, and that the model could explain the anomalous electron recoil excess reported by the Xenon1T collaboration.