In network management, significant research efforts have been carried out to automate and facilitate the tasks conducted by network administrators. However, so far, none of these efforts has exploited the opportunities of jointly using the Situation Management discipline and the mashup technology for network management. This paper introduces an approach, called Rich Dynamic Mashments, to facilitate the daily work of network administrators when dealing with unexpected, dynamic, and heterogeneous situations. We have referred to as nmsits to such type of network management situations (e.g., a sudden packet loss in a core router of a network backbone and an unforeseen slowness in data transmission over a link between virtual routers) and mashments to tunable mashups that use Situation Management for conducting network management tasks. The proposed approach is made up by the models of nmsits and mashments, the mechanisms to automatically recognize nmsitsand dynamically compose mashments, and the architecture supporting these models and mechanisms. We further implement a prototype of our approach and conduct an extensive analysis on networks based on the Software-Defined Networking paradigm. The analysis results have provided directions and evidence that corroborate the feasibility of using Rich Dynamic Mashments as an effective approach for network management in terms of time-recognition, time-composition, time-consuming, time-response, and network traffic.