Hey @Loris_Bogo,
That is some good troubleshooting there and at least we know there is a good chance it isn’t a fault with n8n. I would maybe start with any local policies or installed application differences, Maybe some AV is trying to do content inspection and breaking something.
Hi @Jon,
I’ve managed to make it working by switching to https instead of http, which anyway was a thing I had to do sooner or later.
Apparently this change made Chrome (or windows/AV/whatsoever) regain trust and allow to successfully connect to N8n.
I hope this solves the problem permanently.
Thanks again for your help!
I have the same issue where the workflows do execute, but the spinning circles on each node would not show, the circle on the “Execute Workflow” button would not stop spinning, and the circle on the stop button (when clicked) would not stop spinning.
I am running v0.164.0 in a Docker container from n8nio/n8n on Debian 11. The container is proxied by an Nginx Proxy Manager container.
My findings:
My suspicion is that a browser plugin or some policy pushed to my wife’s corporate laptops is the source of her problem. This would also appear true for @Loris_Bogo in post 20 where his issues only exist using his office workstation, but not from his home workstation over the VPN.
Hopefully this helps with the troubleshooting even though I couldn’t fine the exact problem.
Hey @anon3028993,
That is interesting, So during your tests from your Mint Desktop and Win 10 VM was that just using HTTP? I recently started using NGINX Proxy Manager for my local instance as it is makes it a lot quicker to get a service up and running than manually having to do it.
A browser plugin would be a bit odd but there are a few that work across all 3 of those browsers, It could just be a proxy on the work laptop assuming there is a VPN in place.
This is going to be one of those issues where it is going to be tricky to track it down as we have a mix of HTTP and HTTPS in this thread although your issue and the one before are both solved by swapping to HTTPS.
Correct, my tests from my Mint desktop and the Windows 10 VM worked fine using HTTP.
I agree that a browser plugin is most likely not the source of the issues here. It may be more likely something IP stack related (controlled via a Windows GPO?). Because HTTPS prevents the interception of traffic between the browser and n8n, the HTML/JS cannot be easily tampered with. This might explain why enabling SSL “fixes the problem”.
Unrelated NPM note: Using NPM makes container access much easier than just using unique ports via Docker. My NPM container is using a macvlan network so it has a static IP address on the same subnet as my Docker host. The IP address is in DNS as “proxy”. Every container I deploy is on a user created bridge network that NPM is also connected to. This let’s me spin up containers, setup a “Proxy Host” in NPM, and add a CNAME in DNS for the app, like: “gitea CNAME proxy”. This setup also allows me to move the NPM container (and it’s proxied containers) to another Docker host without breaking DNS as the NPM container keeps it’s IP address regardless of which Docker host it’s on.
I guess the HTTPS side depends on what they are doing, A lot companies still do content inspection with HTTPS traffic which can be a pain and break things. I did wonder if maybe they are using something a bit more basic though and that is blocking some of the javascript or something.
NPM Note: I kind of cheated with mine and just set a couple of catch all subdomains and I just use those so like *. manage and *.tools that point to my NPM IP then I can just make things on the fly without having to worry about DNS. Having user created bridge networks is a good idea, I have got to a point where I have to check what ports I am using before I make an image so I may have to set that up at some point.
I am experiencing the exact same problem on my corporate PC where he circle on the “Execute Workflow” button would not stop spinning, and the circle on the stop button (when clicked) would not stop spinning.
I am self hosting n8n behind a Traefik Reverse Proxy with the n8n ingress being HTTPS. However the actual n8n container I believe is HTTP only at port 5678. I am thinking this could be somehow related to the problem but not 100% sure as most other containers I run are done in similar way.
Hey @Alex_Tzonkov,
Do you know what policies are in place on the corporate laptop? It could potentially be locally installed software trying to be clever and blocking the calls or a corporate proxy.
I am having the exact same issue. I tried a couple of simple workflows and it keeps spinning after clicking execute…
Hey @journeyer,
Welcome to the community 
Could you create a new post for your issue and include how you have n8n setup and if any proxy is being used.
This is a docker-compose setup (following Docker - n8n Documentation) with postgres. This is behind nginx proxy.
I think I might have found the issue for me. For nginx, I need to turn off proxy buffering.
proxy_buffering off;
When I placed this line in the section which has the proxy_pass line and it is now showing the execution status!
Thanks!
