Connect ENETUNREACH 'problem in node 'http request'

Questions
1

Hi all.

I self-host n8n on an old workstation pc at a small business. We use n8n to receive webhooks from our ERP service and send notifications via teams. I have recently encountered the ’ Problem in node ‘HTTP Request‘ connect ENETUNREACH 2603:1036:304:81c::2:443 - Local (:::0)’

Additionally, I use cloudflare to protect the domain and all traffic filters through them. The error only occurs at the HTTP request stage, and sends a post request to our ‘incoming webhook’ plugin for Teams.

This only started recently (about a week ago) and has been persistent ever since. It will error about 1 time out of 10 requests. I have done the following:

  • Forced ipv4 in the env settings on docker
  • Updated to the most recent version of n8n
  • Rebuilt the container post making the changes

Any help would be appreciated!! thanks

2

  1. Check if Docker is using IPv6 and if it’s not necessary, disable IPv6 completely in both Docker and the host operating system.

  2. Disable IPv6 at the operating system level (Linux)

  3. If you’re using Cloudflare, check if you’re blocking IPv6 traffic or redirecting something incorrectly. Sometimes connecting from a non-IPv6 server to an endpoint that responds with IPv6 can cause intermittent errors.

3

Erick,

Thanks for the response. I have completed your line number 1, 2.

Looking into the third one.

4

Unfortunately, none of those answers fixed my issue. It will error out 1/10 times. The weird thing is, I’ve narrowed it down to just one workflow erroring out. For example, one workflow ran about 20 times with no errors (to the same ERP system I’m working with) and my other workflow errored 3 times of 6 runs today.

5

I switched over our windows server to linux and haven’t seen this issue yet