HTTP node in version 1.121.3

Questions
1

In the latest version (1.121.3) of self-hosted, HTTP nodes started timing out.
I tested this by running the same request several times in a row.

It executed several times, then on the 3rd or 4th attempt, it went into a waiting state and get timeout error. In Postman, I run the request 100 times, and there’s no such problem. I tested Curl on the server where n8n is deployed, and this problem doesn’t exist. The server consistently responds.
However, the HTTP node gets stuck at some point and terminates with an error.

Check this out n8n http Node — Screen Studio

2

This appears to be a connection pooling issue in the HTTP node. Try these steps:

1. In your HTTP node settings, enable “Keep Alive” if it’s disabled

2. Set “Max Sockets” to a higher value (default is 50)

3. Alternatively, add `Connection: close` to your request headers to force new connections

If that doesn’t help, you might want to check your n8n logs for more detailed timeout errors. The issue seems specific to how n8n manages HTTP connections versus direct tools like curl.