Combined studies of variable stars and stellar clusters open great horizons, and they allow us to improve our understanding of stellar cluster formation and stellar evolution. In that prospect, the Gaia mission will provide astrometric, photometric, and spectroscopic data for about one billion stars of the Milky Way. This will represent a major census of stellar clusters, and it will drastically increase the number of known variable stars. In particular, the peculiar Gaia scanning law offers the opportunity to investigate the rather unexplored domain of short timescale variability (from tens of seconds to a dozen of hours), bringing invaluable clues to the fields of stellar physics and stellar aggregates. We assess the Gaia capabilities in terms of short timescale variability detection, using extensive light-curve simulations for various variable object types. We show that Gaia can detect periodic variability phenomena with amplitude variations larger than a few millimagnitudes. Additionally, we plan to perform subsequent follow-up of variables stars detected in clusters by Gaia to better characterize them. Hence, we develop a pipeline for the analysis of high cadence photometry from ground-based telescopes such as the 1.2m Euler telescope (La Silla, Chile) and the 1.2m Mercator telescope (La Palma, Canary Islands).