In this paper we present a dataset containing both the network of the "follow" relationships and its growth in terms of new connections and users, all which we obtained by mining the decentralized online social network named Mastodon. The dataset is combined with usage statistics and meta-data (geographical location and allowed topics) about the servers comprising the platform's architecture. These server are called instances. The paper also analyzes the overall structure of the Mastodon social network, focusing on its diversity w.r.t. other commercial microblogging platforms such as Twitter. Finally, we investigate how the instance-like paradigm influences the connections among the users. The newest and fastest-growing microblogging platform, Mastodon is set to become a valid alternative to established platforms like Twitter. The interest in Mastodon is mainly motivated as follows: a) the platform adopts an advertisement and recommendation-free business model; b) the decentralized architecture makes it possible to shift the control over user contents and data from the platform to the users; c) it adopts a community-like paradigm from both user and architecture viewpoints. In fact, Mastodon is composed of interconnected communities, placed on different servers; in addition, each single instance, with specific topics and languages, is independently owned and moderated. The released dataset paves the way to a number of research activities, which range from classic social network analysis to the modeling of social network dynamics and platform adoption in the early stage of the service. This data would also enable community detection validation since each instance hinges on specific topics and, lastly, the study of the interplay between the physical architecture of the platform and the social network it supports.