The ever-growing proliferation of wireless devices and technologies used for Internet of Things (IoT) applications, such as patient monitoring, military surveillance, and industrial automation and control, has created an increasing need for methods and tools for connectivity prediction, information flow monitoring, and failure analysis to increase the dependability of the wireless network. Indeed, in a safety-critical Industrial IoT (IIoT) setting, such as a smart factory, harsh signal propagation conditions combined with interference from coexisting radio technologies operating in the same frequency band may lead to poor network performance or even application failures despite precautionary measures. Analyzing and troubleshooting such failures on a large scale is often difficult and time-consuming. In this paper, we share our experience in troubleshooting coexistence problems in operational IIoT networks by reporting on examples that show the possible hurdles in carrying out failure analysis. Our experience motivates the need for a user-friendly, automated failure analysis system, and we outline an architecture of such system that allows to observe multiple communication standards and unknown sources of interference.