NeophyteI recently had a similar request, in that the issue at hand was to look at (hundreds of) Cisco routers defined in our instance of LogicMonitor and perform the comparison of Throughput between the Primary Interface and backup VLAN interface. The approach I took was slightly different. Problem was the metrics between polls for InOctets and OutOctets metrics differed wildly between the polling intervals (by default 2 minutes). So computing Total throughput in “TotalBps” ((InOctets*8) + (OutOctets*8)) based on these metrics would result in peaks and valleys and this is regardless how these 2 metrics are defined (counter, derive, gauge). So I defined a virtual Datapoint “Total95” in one of the graphs “percent(TotalBps,95)” which returns the 95 percentile value of all available values of datapoint “TotalBps” for the displayed time range, in my case Last Hour, using aggregated data. Now the comparison can be done. Also, the fact you can plot all these devices metrics in Dashboard and save said Dashboard into a Report which then can be fetched programmatically via API which then can be parsed out for individual device analysis is awesome. Instead of going to hundreds of devices one by one and risking exceeding API poll limit you just make one API call to get the report and then crunch the numbers using the scripting of your choice. I am not sure if this is the right approach, of course, but it seems to be working so far, the comparison is yielding correct results, as confirmed by Network Engineers on prem.
