In Pervasive Computing environments, people are surrounded by a lot of embedded services. Since pervasive devices, such as mobile devices, have become a key part of our everyday life, they enable users to always be connected to the environment, making demands on one of the most valuable resources of users: human attention. A challenge of the mobile computing systems is regulating the request for users' attention. In other words, service interactions should behave in a considerate manner by taking into account the degree to which each service intrudes on the user's mind (i.e., the degree of obtrusiveness). The main goal of this paper is to introduce self-adaptive capabilities in mobile computing systems in order to provide non-disturbing interactions. We achieve this by means of an software infrastructure that automatically adapts the service interaction obtrusiveness according to the user's context. This infrastructure works from a set of high-level models that define the unobtrusive adaptation behavior and its implication with the interaction resources in a technology-independent way. Our infrastructure has been validated through several experiments to assess its correctness, performance, and the achieved user experience through a user study.