Http GET node: aborted at TLSSocket.socketCloseListener

Questions
, ,
1

Hi, I have used n8n for half a year in production. Works great!

However, I have recurring problem that I believe I have concluded to be in the http-node.
The same GET-request with cURL works without problems. I’m querying a Mikrotik firewall with RouterOS 7.17.2.

The problem occurs when the payload contains roughly 1000 records or more.

I was thinking about the environment setting N8N_PAYLOAD_SIZE_MAX but it turned out to be 16MB default and the data (recieved by cURL) shows 360 kB so nowhere near the limit…

Any ideas?
Thanks in advance

What is the error message (if any)?

Error: aborted at TLSSocket.socketCloseListener (node:_http_client:464:19) at TLSSocket.emit (node:events:530:35) at node:net:343:12 at TCP.done (node:_tls_wrap:648:7)

Please share your workflow

Information on your n8n setup

  • n8n version: 1.76.3 (Self Hosted)
  • Database (default: SQLite): SQLite
  • n8n EXECUTIONS_PROCESS setting (default: own, main): main
  • Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
  • Operating system: Ubuntu 22.04
3

Hey @Jonas_Carlsson,

Have you tried increasing the timeout? Sometimes it’s as simple as that.

You could also try a similar request with Postman to see what happens and get more detailed information from the request.

4

Hi! Thanks for your suggestion. Yes, I have. I have set it to 300000 ms (5 mins). The ABORT-message comes immediately. I also have the retry set to 5 times with 5000 ms between.

The strange thing is that when I take the simple GET-request and run a cURL manually (outside of n8n) I get the response immediately.

5

Let’s see if I understand:

The same GET request, with the same payload, works from your server if you execute it from the command line.

But when running in n8n the problem occurs.

Is that exactly what is happening or didn’t you try to run the same request using the command line?

6

Thanks for your interest in this, much appreciated!

When I had the issue the last time ran the HTTP-node with the GET-request manually and got ABORTED.
On another computer, outside of the docker world (where the n8n resides) I ran a cURL-GET with the same parameters and boom, I got all 2000 records in 2 seconds.

The next time I see it, it happens several times a week (I just have to be attentative) I can record/screenshot the node and the cURL-command with its result.

7

I see. Try to make the cURL request from inside the docker container.
It may be related to the container or server settings, instead of n8n.