Research has shown that the N-1 criterion is not adequate to capture the uncertainty faced by system operators due to (i) the random occurrence of contingencies and (ii) the forecast errors in load and production of renewable energy sources. Instead of the N-1 criterion, security management approaches based on probabilistic methods have been proposed. We contribute to these approaches by developing methods to compute and control the probability that the system will experience voltage stability issues. More specifically, we developed the following methods: (i) a method to approximate the probability of voltage instability, (ii) a method to speed up Monte-Carlo simulations to estimate the probability of voltage instability and (iii) a chance-constrained optimal power flow formulation that ensures that the probability of voltage instability is less than a pre-defined threshold. All methods consider probabilistic forecasts for load and wind power as well as the probability of occurrence of contingencies.