One of the main enablers of Distributed systems, edge computing transacts nearer to the end consumers and provide minimal latency production services. This is due to the fact that; edge nodes have diverse deployment environments, constrained resources as well as sensitivity to software and hardware failures. This work addresses the issue of applying complex machine learning techniques to predict when certain edge computing nodes might fail. As datasets, this work employs both operational variables, ambient factors, and failure logs that have occurred in the past to develop a prediction model that could probably alert on possible issues. In other words, such significant variables as working memory, CPU load, network traffic, and environment to identify trends and precursors of system imbalance are analyzed. The use of the suggested solution reduces operating cost and time hence increase the reliability of edge computing systems by providing proactive maintenance. This work confirms that predictive analytics can enhance the edge infrastructure so that it will promote the distributed computing systems that are tougher and better.