In response to European sustainability directives, a rapid growth of smart sustainable buildings (SSBs) is expected in the coming years, particularly in low voltage (LV) networks. This opens up new pathways towards exploiting SSB-driven flexibility for the provision of upstream ancillary services (AS). However, tapping into these resources is contingent on properly dealing with the inherent end-user uncertainty. In that light, this work proposes a detailed mixed-integer nonlinear, joint chanceconstrained optimization framework that coordinates SSBs in order to provide voltage and congestion support AS at the MV-LV interface. The sources of intractability (AS model, network nonlinearity, uncertainty) are linearly re-formulated, comprising a mixed-integer linear optimization model. Building on previous work, the proposed operational planning framework limits the occurrence of technical violations, so as to allow the LV network to dedicate all available resources exclusively towards providing AS support to the MV network, in real-time. The model's validity and performance are evaluated on a real Luxembourgian LV grid, assumed to host a significant number of residential SSBs.