Update to this thread…
I have been mulling this question over since posting the flow and while I still don’t have a complete answer, I wanted to give the customer a better answer than the example flow.
Reaching out to them and talking over their application more, it became very clear that something really was amiss in Node-RED, so they shared their flows with me.
Once I put their flows into a RIO here in the building, I was able to run a Wireshark network capture.
Digging into it showed a surprising result…
With Node-RED disabled via groov Manage, there was almost zero network traffic, just the odd NTP time sync packet.
With Node-RED enabled, there are a lot of chatter over the wire.
Looking at their flow, I reconized all the nodes in use, bar one… A MongoDB node.
Removing this node from the flow and doing a deploy stopped 99% of the network traffic instantly.
The really odd thing is that the node generated traffic even when it was not in use.
With no msg.payloads going in or out, the node should not have been generating any network traffic. At least that’s what I expected to happen (or not happen as the case may be).
And yet it was.
To the tune of 33Kb per minute.
No wonder the customer cell modem data use was higher than expected.
We have raised an issue on the nodes developer github. Wait and see.