Thank you - I will try it today or tomorrow evening.
I am very experience with Java and I like the possibilities there to inspect and profile the Java Virtual Machine. I can very easily detect which code "burns" the CPU usage.
It would help me, if i find out, which node causes the problem.
Regarding the feedback loops or synchronizing problems: Most flows are very linear in means of reading data from a source (e.g. modbus or an external API) and writing it to an InfluxDB v2. I am using 3 or4 global context variables to combine and calculate values based from two different flows. But those are triggered by timer and not events. I don't think I have feedback loops in the manner of MQTT or similar.
BTW: I have these problems now since months, even the upgrade to Node-RED 2.x did not resolve it (currently using latest-12).
