Monitoring resource availability along Cloud-IoT networks in a lightweight and fault-resilient manner is a challenging research problem due to scarce resource availability, infrastructure dynamics, and platform heterogeneity. In this article, we illustrate a thorough experimental assessment of a self-organising and fault-tolerant monitoring service, FogMon, especially targeting Cloud-IoT settings and capable of probing hardware resource, latency and bandwidth. The assessment is carried out over networks made up of 20 to 40 nodes across two testbeds within the Fed4Fire+ federation. As a result of the assessment, we agilely improved and refined FogMon into FogMon 2, which settles at TRL5 and improves on monitoring accuracy and fault-resiliency. Experimental results show how FogMon 2 can promptly and suitably handle different types of infrastructure failure, with an average relative error of 10% on measurements and limited footprint on hardware and network resources.