No error apears, just a 499 when the connection is closed, the data doesn’t even reach the workflow, there’s no execution.
Redeplyoing the container seems to fix it for a few hours, but it starts doing the same thing after that.
Bellow is the http log from railway, the 200s are JSON request that work fine, the 499 are XML requests sent by the SalesForce, it timesout after 1 minute, but if you try it with postman, it hangs indefinitely
Based on the GitHub issues and your description, this appears to be a known problem with n8n’s HTTP Request node hanging when processing XML data. Here are some practical solutions to try:
**Immediate fixes:**
• Set a specific timeout value (like 30000ms) in your webhook node’s options
• Try enabling “Retry on Fail” in the node settings
• Add request headers explicitly: `Content-Type: application/xml` or `text/xml`
**Configuration adjustments:**
• Increase your Docker container’s memory allocation if possible
• Consider setting `N8N_PAYLOAD_SIZE_MAX` environment variable to handle larger XML payloads
• Try switching the execution process mode from “own” to “main” in your n8n settings
The fact that redeploying temporarily fixes it suggests a memory leak or connection pooling issue. As mentioned in the [github.com](http request hangs · Issue #16738 · n8n-io/n8n · GitHub) issues, this is an intermittent problem that the n8n team is aware of, particularly with batch processing and slow external services.
