Because the frequency and character of rifting events along mid‐ocean ridges are largely unknown, how the repetition of such events gives rise to rift structures is unexplored. The Asal rift in the Afar depression of Djibouti, Africa, provides the world's best subaerial analogue for young slow spreading mid‐ocean ridges. Seismic, geodetic, and field observations of a seismovolcanic event in 1978 at Asal yield estimates of the fault and dike locations, geometry, displacement, and volume of basalt extruded in a rifting event. A 6–9 kyr‐old lake shore highstand at Asal has been warped downward by 70 m, providing a Holocene measure of the vertical deformation across the rift. The rift topography furnishes an older datum, which we infer to be 34±6 kyr old using the Holocene deformation rate. We find that faults throughout the rift valley are active; Holocene slip rates diminish beyond 4 km from the rift axis; late Quaternary rates decrease beyond 6–7 km. The Holocene slip rates are used to estimate repeat times by taking the displacement on the faults which slipped in 1978 as characteristic; we find tectonic events on individual faults recur every 200– 300 years. Half the rift faults slipped together in the 1978 event. If this is typical, then groups of faults are activated every 100–150 years. We suggest that half the events take place in the rift axis accompanied by volcanic extrusion; the remainder occur peripheral to the neovolcanic zone and involve fault slip only, both events having a repeat time of 200–300 years. Given the 10 km width of the rift and its 16 mm yr −1 spreading rate, the mean age of the material in the rift should be ∼350 kyr, an order of magnitude older than the inferred age of the formation of the rift topography. The subsidence rate of the rift axis during the past 35 kyr is 8–9 mm yr −1 , with the rate of infilling by volcanic extrusion <1 mm yr −1 . The resulting net subsidence rate, about equal to the half‐spreading rate of the rift, could not be sustained for 300 kyr without significant infilling by lavas. Thus both observations suggest that the long‐term vertical deformation in the rift has not been steady state. Instead, we suggest that there is a rifting/filling cycle at Asal, with the most recent filling episode ending ∼35 kyr.