When executing my workflow, the HTTP node completes but doesn’t move onto the next node (image 1). Interestingly, the input/output viewer bugs during this for maybe 10-20 seconds before returning to normal (image 2).
Sounds like the HTTP node is finishing but not handing off data properly. A couple of things you can try:
Check the binary data handling → sometimes the next node expects JSON instead of binary. You may need to use a Move Binary Data node to convert before passing it along.
Since you’re on Docker + Apache, also check reverse proxy/WebSocket settings (timeouts or buffering can cause the [WebSocketClient] Connection lost, code=1006 issue).
Make sure you’re on the latest n8n version, sometimes these bugs are fixed in newer builds.
If you want, feel free to share a simplified workflow or reach out to me at [email protected] and I can help debug further.
Thanks for your reply. I can’t do the former as I need to upload the file in the HTTP request, which wouldn’t work with the JSON extract. In regard to the latter, here is my env file/Apache config if that helps:
Apache
<VirtualHost *:443>
ServerName n8n.<mydomain>.com
SSLEngine on
SSLCertificateFile /etc/letsencrypt/live/n8n.<mydomain>.com/fullchain.pem
SSLCertificateKeyFile /etc/letsencrypt/live/n8n.<mydomain>.com/privkey.pem
ProxyPreserveHost On
ProxyRequests Off
RequestHeader set X-Forwarded-Proto "https"
RewriteEngine On
RewriteCond %{HTTP:Upgrade} =websocket [NC]
RewriteRule /(.*) ws://127.0.0.1:5678/$1 [P,L]
ProxyPass / http://127.0.0.1:5678/ keepalive=On retry=0
ProxyPassReverse / http://127.0.0.1:5678/
ProxyTimeout 3600
TimeOut 3600
LimitRequestBody 0
ErrorLog ${APACHE_LOG_DIR}/n8n_error.log
CustomLog ${APACHE_LOG_DIR}/n8n_access.log combined
</VirtualHost>
I attempted to add:
N8N_PAYLOAD_SIZE_MAX=1073741824
N8N_DEFAULT_BINARY_DATA_MODE=filesystem
Which would in theory increase the limit more than enough for my 12mb upload and 170mb output. Unfortunately, this did not work. To clarify your last point, I am on the latest version.