Workflow Editor Blank (White) After Upgrading

We are running n8n in a Docker container and were a few versions behind (last upgrade was in December). After performing a pull and re-running the docker containers, everything appears to be running on the server correctly, however, when attempting to access the workflow editor via a browser, we only get a blank page. The page title shows - Workflow Automation, but nothing else populates.

I tried creating a clean SQL Lite Database, but the page was still blank. Nothing out of the ordinary appear in the consoles for Traefik or n8n in either case.

Hey @lits welcome to community!

Please provide

  • screenshot how it looks like
  • docker configuration (docker-compose.yml file or docker run command depended on what you use)
  • What version do you use now and what did you use before

This will help us help you will issue

1 Like

So did a bit more digging - I apologize, I don’t really remember what version we were upgrading from, but after doing a bit of troubleshooting, I can get n8n back up and running if I pull version 0.155.2. The moment I pull version 0.156.0, the blank screen starts.

My docker-compose.yml file is/was identical to what is in Github: n8n/docker-compose.yml at master · n8n-io/n8n · GitHub – the only modifications that I have done recently was to change it from using the latest tag.

Hi @lits, this sounds odd. Are your browser and n8n running on the same machine? If not, are you able to open the UI on the machine running n8n (if it’s a server without a browser, you could alternatively try curl -v http://localhost:5678)?

Hi @MutedJam - n8n is running as a docker container on a Linux server. The n8n server is also behind a traefik container? I’m not super familiar with it, but it is based upon the docker-compose file listed in the instructions.

What’s really odd is if you look, the page title gets returned, but the page contents is completely blank.

I just started having this exact error after a recent upgrade. Running n8n behind traefik as well. What’s interesting is that I can access the instance using the Tunnel URL.

Hi @physx911, I was unfortunately not able to reproduce this problem based on the information available so far. My best guess is that the reverse proxy server or load balancer isn’t routing all traffic as it should (using the tunnel URL would circumvent your reverse proxy). Seeing this is a rather old thread, perhaps you could open a new topic and provide all information required to reproduce your problem?

I also have the same issue… it’s happened since the update

I’m using cloudfalred tunnel as a reverse proxy for all other applications works fine with the reverse proxy
just the n8n creating issues…

@physx911 do u have any community nodes? maybe we have a node that’s causing this issue
also type of storage are u using? (eg MySQL, PostgreSQL)

@thefourCraft what version are you on and what did you upgrade from?

If you access the instance without the tunnel does it work?

Hi @Jon
I can access it without the tunnel. I can also access it from the webhook API
also, I can access it if I open another browser (from the regular URL)

I am using docker to host
my version is 0.218.0

Hey @thefourCraft,

That seems to rule out a lot of things and if I am reading it correctly it sounds like it is an issue with one browser so probably wouldn’t be linked to community nodes.

Which browser is having the issue and assuming it has a private browsing mode does it work from there? My thoughts at the moment is it is some kind of cache issue.

I do not have any community node installed. However, the service is served via Cloudflare CDN proxy, perhaps as Jon mentioned maybe a cause? The issue appears and disappears on its own though. If the problem does appear, it appears on Chrome and Firefox including their respective incognito modes.

Are you using any of the caching features? There was someone a couple of weeks ago using one of the minify / cache options and it was causing an issue. As a test I would maybe still use cloudflare but turn off the cache options.

I will take a look at this because this problem didn’t exist before
it happened only recently, also disabling Cloudflare cash isn’t possible for a lot of users
I will create a rule for that domain to test it. Not sure what’s happening…

In your case it is a bit different as you are using the tunnel option so that could add more problems to it but from your testing it looks like n8n is working as expected apart from the one browser which to me would suggest it is something with the browser.

Looks like this is the issue

Uncaught SyntaxError: Invalid or unexpected token
DevTools failed to load source map: Could not load content for chrome

why dose it think that JavaScript isn’t enabled?

If I open the page in other window it will work (other google Chrome profile)

The other chrome profile is an interesting thing so it is now localised to your chrome profile which would suggest an extension or setting you are using. Do you have an adblocker or maybe in browser vpn extension?

here are my extensions, but I don’t think they are the problem

They all look ok to me but your testing kind of points to there being an issue there, if it was n8n or the tunnel I would expect it all to fail not just one profile.

It is an odd one, have you tried disabling them all and seeing if it works?

I did disable all of them no change, hers a side by side photo with other window open

it’s the same website, different chrome profiles (google profiles) and different users (n8n users)
I did a hard reset on cookies on both…
also cloudflare all cashing rules are of and the tunnel doesn’t have any additional changes
the only rules that are in place are WAF rules…

this didn’t happen before, n8n worked without any issues on cloudfalre…