A network state tracking system tracks the state of a network and recommends an action when the state of the network changes. The network state tracking system detects one or more nodes on the network and generates one or more snapshots of the network. The snapshots are generated by obtaining first and second status data from the nodes and detecting whether a change in the network has occurred based on the status data. The snapshots are generated based on the first and second status data when a change in the network is detected. The snapshots are then used to train a machine learning model to recommend an action to take when a change in the network occurs.