To ensure the dependability of a Multi-Agent System it is essential to enable it to detect and handle critical states of its agents. This can be done by either enabling the agents to deal with unexpected situations or by adding a fault-management component to the platform. This paper has two aims. Firstly it is an attempt to motivate the need for a generic fault-management service offered by the platform, since we believe that the current lack of deployed MAS is mainly a result of their brittleness. Secondly, we want to present an open platform-independent framework based on event-streams and report the evaluation results.