Echo-chambers present a persistent issue in communication networks, and especially in centralized ones as existing work tells us. The dynamics of this phenomenon is scarcely studied for decentralized networks. Thus, our motivation to investigate factors of echo-chamber formation in these networks, and in particular in dedicated instances in Mastodon as such. Utilizing community detection and adapted Shannon entropy metrics to analyze instance diversity within communities we can not conclude that dedicated instances present strong influential factors for echo-chamber creation in Mastodon. Our results show that while the instance structure affects both where followers originate and the instance diversity within the follower networks, users boost toots from a highly diverse variety of remote instances indicating that content easily crosses instance boundaries.