ECONNRESET on AI Agent

Questions
,
1

Lately, I’ve been seeing a bunch of “read ECONNRESET” errors on the AI Agent, which has an easy fix by just retrying the request. However, that requires me to have an if node set up behind every AI Agent, which makes the workflow quite bulky. I’m also not sure if this error is linked to an n8n update or if something degrading performance is going on at OpenAI, but this has been going on for about 5 days now, which makes me think they would have caught it by now.

Since I didn’t really see anyone talking about it, I wanted to make this post to see if other people were seeing this aswell

Information on your n8n setup

  • n8n version: 1.78.1
  • Database (default: SQLite): Postgres
  • Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
    And I run n8n in Queue mode
2

So far it did not happen to me and I am not sure if you already tried but I use that with the HTTP node and the agent has it as well. Under Settings Tab you can check “Retry On Fail” with a wait time in between request. Maybe that helps.

3

Retry on Fail is not usable for me because the condition of that is that it has to use the “Stop Workflow on Error” option. If any of the other 2 options are used, it will always output the first execution of the node that errored, even if a retry succeeded.