Not sure about the example and how complete it is supposed to be but the nodes are for example not connected. I guess however that they are in your workflow. That error sounds more like a temporary connection issue, for example problems with resolving the IP address or something else.
Anyway. just to be sure I just tested the following and it worked fine:
Honestly no idea. And actually do not think that it should matter much what is in front of n8n (here cloudflare and nginx) in this case. As it is an outgoing connecting and so should not matter. Unless I understand something wrong.
Then the question is probably why it got added in the first place? Because by default it is not there. Then it makes also total sense. If 2-3 MB have to be downloaded and it takes longer than 10 seconds (which is the timeout you had apparently set) and it times out, then it did exactly what it was supposed to do, or do I understand something wrong?
I set timeout to single step that downloads file. There was no problem with downloading file (timeout was not reached). But following steps failed with ESOCKETTIMEDOUT error. Also they failed after restarting container with application and running specific IO-related steps without configured timeout.
So just to understand you correctly. You get the timeout on the Dropbox node? But you only get it if you have a timeout set on the HTTP-Request node? If you remove the timeout from that node, the Dropbox node works fine?
I had timeout on “Fetch book” node. There was no ESOCKETTIMEDOUT errors after removing timeout.
As I mentioned earlier I’m not perfectly sure that this things are related.