This paper proposes an energy-reserve market clearing model for microgrids considering probabilistic security criteria. The probabilistic security criteria include pre-selected scenarios associated to unreliability of generators and uncertainties caused by stochastic behavior of loads and renewable units. In contrast to traditional deterministic reserve-constrained market clearing models, this paper determines the optimal amount of reserve as the point at which the sum of its operating costs and the expected cost of load shed reach a minimum. The proposed model is formulated as a two-stage stochastic programming problem, where the first stage represents the hour-ahead energy-reserve market, and the second stage the balancing market. An energy management procedure is developed in order to implement the stochastic programming problem on the real microgrid ATENEA located at the installations of the National Renewable Energy Centre of Spain. The generation and reserve schedules obtained with the stochastic approach are assessed and compared with those of a purely deterministic security-constrained case.