I send out a broadcast every 30 seconds to 3 inverters for status information - and use this to populate a dashboard.
Since upgrading to 2.2.2 and also the latest Buffer Parser i have had problems with the calculation that is used to placed data on the dashboard not working - coming up with a NaaN
As part of trying to track this down (before raising an issue) i have activated some debug nodes - however with a minimum of 3 x responses coming in every 30 seconds - it is hard to track down where the problem is - is there a way to throttle (or freeze) the debug window ?
Also, along the flow (starting at the start and progressing along...) have the debug nodes and process the messages at those points? To reduce the overload: cut the flow to that point and investigate the messages at those points and see if the structure (or what ever) is what you need/expect?
Also, to help with the messages and "who is with whom": use a join node. Then create one message with 3 parts.
That way you see it as a slice from all three points and at the same time.
Another strategy is to set the debug nodes output to Console. Then the debug info goes to the node red log and you can run it for a bit then examine it at your leisure.
These are currently live systems and control the realtime power in our house (either charging from excess solar (not at the moment in Sydney!) or pushing power back out from the batteries to balance grid draw.
So the longer i push out the queries to get the status of the inverters the further they get from the optimum operation.