Recent network system becomes more complex and larger because of virtual LAN and other virtualization technologies. Proliferation of a variety of network switches and routers makes a network system gigantic; hence, management of the misty network faces problems. This is the largest factor that deteriorates stable operation of a network system that should be robust and reliable. One of the promising solutions to keep a network system simple and understandable is introduction of the monitor tools that makes a network system visual and observable. We introduced the sFlow technology in addition to the traditional SNMP-based network node management (NNM) system. We could take statuses of network nodes by NNM such as hardware failure, and also we could grasp long perspective of network traffic at one view by the sFlow. In addition, an integrated log management system was introduced to collect all events on the whole network system. As a result, we could detect a trouble outbreak in real time even if a trouble occurred on the end point of the network, and could solve the problem promptly. We describe a way to achieve an observable network system to maintain stable network operation.