Semeru (Indonesia) is a persistent eruptive volcano, but its volcanic activity remains poorly constrained. Here we used a combination of different remote sensing data to improve the understanding of Semeru. By differencing high resolution (2m) Digital Elevation Models (DEM) derived from optical data (EarthDEMs), we estimated a bulk volume of ~43.5x106 m3 of lavas erupted from 2014 to 2020. Thermal data from MIROVA show multiple peaks in the volcanic radiative power and in the cumulative radiant energy (VRE) related with lava flow extrusions. Time series of deformation, obtained with Synthetic Aperture Radar Interferometry (InSAR) from Sentinel-1 data, show on both orbits negative line-of-sight displacements of the SE flank from 2014 to 2023. We modelled this deformation with a Boundary Element Method, showing that displacements are related to a limited (≤1x106 m3) deflation of a shallow (<2 km) reservoir. We used the volume estimated by EarthDEMs to calibrate the VRE and to calculate the erupted volume of lava from VRE for periods not covered by EarthDEM, allowing in the future the near-real-time estimation of volumes from MIROVA data. Erupted volumes of lava over the time reveal a steady state activity. The large discrepancy between the erupted and the modelled volume suggests that the shallow reservoir is a hydrothermal reservoir or an ephemeral gas-rich magma reservoir, with the persistent volcanic activity fed by deeper reservoir(s). This study highlights the importance of using different remote sensing data to monitor, quantify, and interpret the volcanic activity in poorly monitored and studied volcanoes.