In response to the intricate management demands of network devices and the proliferation of massive log data, this paper proposes a four-tier architecture for a network device log analysis platform. It delves into the functional requirements encompassing data collection, storage, processing, and application, with a focal point on data processing technology implementation. Employing big data techniques, such as k-means cluster analysis and trend analysis, this study showcases the network log situation and forecast results through a network health analysis instance, furnishing a pragmatic solution for intelligent operations and maintenance.