N8n Interface Not Accessible – Even After Full Project Reset on Easypanel

Questions
1

Hi everyone,

I’m facing a very strange and frustrating issue trying to run an n8n instance on Easypanel. I’ve tried everything I could think of — including completely deleting the project and setting up a new one from scratch — but the problem still persists. I’m hoping someone in the community might have gone through something similar and can help.

:jigsaw: What I’m Trying to Do

I’m deploying n8n via Easypanel using this architecture:

  • n8n_start
  • n8n_webhook
  • n8n_worker
  • postgres
  • redis

Everything seems to start correctly. All containers show a status of “Up,” the database and Redis work fine, and the n8n_start logs display this message:
Editor is now accessible via: https://n8n-n8n-start.mybz2e.easypanel.host

However, that URL never loads, and I always get a ERR_CONNECTION_TIMED_OUT.

:test_tube: What I’ve Already Tried

  1. Deleted and recreated the entire project from scratch on Easypanel.
    I nuked the whole stack (project, containers, configs) and set it up again — same issue.
  2. Checked the logs from n8n_start.
    All migrations run successfully, no critical errors appear. But right after the success message, the container receives SIGTERM and shuts down.
  3. Ran curl http://localhost:5678 inside Hostinger’s SSH terminal.
    Got: curl: (7) Failed to connect to localhost port 5678.
  4. File permission warning (.n8n/config):
    I see the message Permissions 0644 for n8n settings file are too wide, but logs say it’s currently just a warning.
  5. Set the environment variable N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true.
    No change in behavior.
  6. Configured SMTP environment variables properly for email recovery.
    Just in case, but this seems unrelated to the core issue.
  7. Double-checked Traefik + domain port mapping.
    Everything is pointing correctly to internal port 5678.

:white_check_mark: What Is Working

  • Redis and Postgres are both up and accessible.
  • n8n migrations complete without errors.
  • Containers spin up and seem fine… for a few seconds.
  • Domain mapping in Easypanel is correctly set to port 5678.
  • n8n_worker and n8n_webhook remain “Up.”

:x: What’s NOT Working

  • I can’t access the n8n interface via the provided domain (https://n8n-n8n-start.mybz2e.easypanel.host).
  • Local curl to port 5678 fails, even though logs say n8n ready on 0.0.0.0, port 5678.
  • n8n_start receives a SIGTERM and shuts down shortly after initializing, seemingly on its own.

:brain: My Best Guess

The n8n_start container initializes correctly, executes all the database migrations, then something is killing it silently — maybe a network config issue, an internal crash not shown in logs, or a proxy/Traefik misconfiguration.

It seems like n8n is being killed before it can actually start the HTTP server properly, or it’s being blocked at the network/proxy layer.

If anyone can offer any advice or guidance, I’d be immensely grateful. I’ve been stuck on this for days, and even after wiping and recreating everything, the issue still persists.

Thank you so much in advance

2

Hi @Freeloader


What do the Environment Variables for these services look like?

3

One possibility is that the service’s healthcheck setup is expecting the /healthz endpoint to be active, but maybe it isn’t.

See: https://docs.n8n.io/hosting/logging-monitoring/monitoring/#healthz-and-healthzreadiness

Just something to check. The SIGTERM sounds like the environment is intentionally killing the process be cause it believes something is not working.

4

hey op you got the solution i am facing the same

5

i got it solved
you need to open port https TCP 443

then everything will work.

thanks