Hey team, I have been using n8n for a long time and the biggest struggle was Websockets with nginx in the past. After we solved that n8n has been running smoothly. Until now, with 1.97.1.
opened 10:49PM - 15 Jun 25 UTC
closed 01:33AM - 16 Jun 25 UTC
in linear
### Bug Description
I installed n8n on a VPS, using docker compose, without Tra… efik since I already have Nginx set up on the VPS.
Since then, I haven't been able to run any workflow, because of a "Connection lost" alert.
I tried to switch to previous versions, but the error is still the same.
NB: I have multiple apps running on the VPS, some of them using web sockets with no problem.
Here are my `docker-compose.yml` and `/etc/nginx/sites-available/n8n.mydomain.com`
```
services:
n8n:
image: docker.n8n.io/n8nio/n8n
restart: always
ports:
- "5678:5678"
environment:
- N8N_HOST=n8n.mydomain.com
- N8N_PORT=5678
- N8N_PROTOCOL=https
- NODE_ENV=production
- WEBHOOK_URL=https://n8n.mydomain.com
- GENERIC_TIMEZONE=Europe/Berlin
- N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true
- N8N_RUNNERS_ENABLED=true
- VUE_APP_URL_BASE_API=https://n8n.mydomain.com
volumes:
- n8n_data:/home/node/.n8n
- ./local-files:/files
volumes:
n8n_data:
```
```
server {
server_name n8n.mydomain.com;
location / {
proxy_pass http://127.0.0.1:5678;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection “connection_upgrade”;
chunked_transfer_encoding off;
proxy_cache_bypass $http_upgrade;
proxy_ssl_server_name on;
}
listen 443 ssl; # managed by Certbot
ssl_certificate /etc/letsencrypt/live/n8n.mydomain.com-0001/fullchain.pem; # managed by Certbot
ssl_certificate_key /etc/letsencrypt/live/n8n.mydomain.com-0001/privkey.pem; # managed by Certbot
include /etc/letsencrypt/options-ssl-nginx.conf; # managed by Certbot
ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem; # managed by Certbot
}
server {
if ($host = n8n.mydomain.com) {
return 301 https://$host$request_uri;
} # managed by Certbot
listen 80;
server_name n8n.mydomain.com;
return 404; # managed by Certbot
}
```
### To Reproduce
1. Run Docker compose without Traefik
2. Login to n8n.yourdomain.com
3. Go to workflow
### Expected behavior
No connection failure.
### Operating System
Debian 12
### n8n Version
1.97.1
### Node.js Version
20.19.2
### Database
SQLite (default)
### Execution mode
main (default)
opened 03:14PM - 26 May 25 UTC
closed 09:21AM - 30 May 25 UTC
in linear
### Bug Description
When upgrading to version 1.95.0 the editor does not work p… roperly anymore. It shows "lost connection". Thus it is not possible to execute workflows in test mode. The rest of the app seems to work fine though.
Log Output:
```
n8n-02-n8n | 2025-05-26T15:01:14.599Z | debug | Skipped browserId check on /rest/push {"file":"auth.service.js","function":"resolveJwt"}
n8n-02-n8n | 2025-05-26T15:01:14.601Z | warn | Origin header does NOT match the expected origin. (Origin: "https://n8n.mydomain.com", Expected: "wss://n8n.mydomain.com") {"scopes":["push"],"expectedOrigin":"wss://n8n.mydomain.com","headers":{"host":"n8n.mydomain.com","origin":"https://n8n.mydomain.com","x-forwarded-proto":"wss","x-forwarded-host":"n8n.mydomain.com"},"file":"index.js","function":"handleRequest"}
```
### To Reproduce
- Update to version 1.94.0 to 1.95.0
- Open any workflow
### Expected behavior
Editor working again without showing connection error in top right corner.
### Operating System
Ubuntu Server 22.04 LTS
### n8n Version
1.95.0
### Node.js Version
n/A
### Database
PostgreSQL
### Execution mode
main (default)
Even downgrading back to 1.94* did the job
UPDATED:
The behavior is so strage, when loading n8n it gives an login error, then I authenticate and it redirect me and loads the flows. Yet if I refresh again I am not logged in.browserId check failed on /rest/login
UPDATED 2:
miko
June 19, 2025, 1:57am
2
Same issue here with ‘Origin header does NOT match the expected origin. (Origin: “foo.com:25678 ”, Expected: “foo.com ”)’ im using behind cosmos cloud. it just work only rolling back to 1.94.1
On my case I had to restart chrome 2 times.
1 Like
system
June 27, 2025, 8:11am
4
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.
I am stuck at the moment.