The publish-subscribe (pub-sub) paradigm is maturing and integrating into community-oriented collaborative applications. Because of this, pub-sub systems are faced with an event stream that may potentially contain large numbers of redundant and partial messages. Most pub-sub systems view partial and redundant messages as unique, which wastes resources not only at routers, but also at possibly resource constrained subscribers. In this paper, we present Caeva, a customizable and adaptive event aggregation framework. The design of Caeva exhibits three novel features. First, the tasks of merging messages and eliminating redundancies are shared among multiple, physically distributed brokers called aggregators. Second, we design a decentralized aggregator placement scheme that continuously adapts to decrease messaging overheads in the face of changing event publishing patterns. Third, we allow subscribers to choose a notification schedule that meets their specific needs. Results of extensive experiments show that Caeva is quite effective in providing flexibility and efficiency.