Diagnosing a problem in an SDN is still a very hard task. Most of the network administrators are stuck with tools such as ping, traceroute and SNMP for checking connection and device status, or more passive tools such as tcpdump to examine network traffic. In this poster we will present some of the most common failure factors on an SDN-based network: race conditions; controller software bug, equipment protocol implementation. We will introduce the Kytos SDN Platform and the work we are currently doing to implement a debugging tool based on it that deals with this failure factors.