Reactive power planning has always been a key research area in power distribution engineering; technically and economically. However, the problem needs to be revisited to consider several aspects of modern distribution systems, such as a high penetration level of renewable resources with intermittent nature; the microgrid concept and the possibility of system operation in grid-connected or isolated single microgrid mode, or isolated multiple microgrids; and probabilistic or hourly load profile. Motivated by these needs and considering all these aspects, this paper presents a generalized approach for probabilistic optimal reactive power planning in modern distribution systems. A new index is defined to probabilistically assess the success of microgrids in terms of real and reactive power adequacy and voltage limit constraints. Afterward, the reactive power planning is performed to reduce the annual energy losses of the grid-connected system and increase the defined microgrid success index. The problem formulation and solution algorithms are presented in this paper. The well-known PG&E 69-bus distribution system is selected as a test case, and through several sensitivity studies, the effect of optimization coefficients on the design and the robustness of the algorithm are investigated. A cost-benefit case study is also presented to determine the optimum total size of distributed reactive sources for the system under study.