Gravitational redshift is thought to be one of the classical tests of general relativity. It is a direct experimental method of testing metric-based nature of gravity and the equivalence principle. The effect of gravitational redshift has been detected in a number of ground- and space-based experiments as well as in observations of compact objects, in full agreement with the predictions of general relativity. During my talk, I will show that the same effect is observed on the scales of galaxy clusters -- the largest and most massive gravitationally bound objects in the Universe. In particular, I will discuss how gravitational redshift manifests itself in velocity distributions of galaxies in clusters, how it can be inferred from the data and compared with the predictions of general relativity. The focus will be on the first detection of the effect based on the analysis of spectroscopic redshifts from SDSS.