Workflow into production webhook error

[tlpui]

Hello,
I am experiencing an issue with a workflow that uses a Gmail Trigger and gmail node: send message and wait for response.
When I was testing the workflow manually from the editor (clicking on the button execute workflow), the feedback link (where I can put a text giving feedback on the response) was correctly generated with the public URL of my n8n instance deployed on Railway (https://......railway.app). In this scenario, the workflow functioned without any problems.
However, when I activate the workflow to run automatically (in production mode), the link sent via email for giving feedback always starts with http://localhost:5678/.... This prevents the workflow from continuing, as the URL is not accessible from outside the local machine.
I have verified that the WEBHOOK_URL environment variable is correctly configured in my Railway deployment with the public URL of the instance. . Despite this, the problem persists.
Could you please help me understand why the active workflow is not using the WEBHOOK_URL variable and is instead generating the localhost URL?
Thank you for your assistance.
Best regards

[zee.choudhry]

Hello @tlpui

Could you share your docker command where you are using the WEBHOOK_URL environment variable

Below you can find the example command which should work OK:

docker run -it --rm --name n8n -p 5678:5678 -e WEBHOOK_URL=https://mycustomurl.com/ -v n8n_data:/home/node/.n8n docket.n8n.io/n8nio/n8n

Notice https:// and trailing / (slash)

Can you also share a simpler workflow for Gmail Trigger and Gmail node for : send message and wait for response that can reproduce the issue?

[tlpui]

Hello,

Thank you for your response and the Docker command example. I am deploying my n8n instance using a Docker container on Railway. Therefore, I cannot directly execute the docker run the command you provided me.

As you can see in the attached screenshot, I have configured the WEBHOOK_URL environment variable with my public Railway instance address (https://...railway.app), just as you advised.

Despite this variable being correctly set, the problem persists:

• In test mode (within the editor): The webhook link is generated correctly using the Railway URL.

• In production mode (active workflow): The webhook link sent via email always starts with http://localhost:5678.

To rule out any issues with my original workflow, I have created a much simpler one that still shows the same error in production.

I hope we can somehow reach to a solution.

Thank you for your time and help.

Best regards

[tlpui]

I have recently tried to add the variables as well. Nevertheless, the result is still the same: it shows ‘localhost’ when triggered automatically, but the normal URL when done manually.

N8N_HOST = .....railway.app

N8N_PROTOCOL = https

N8N_EDITOR_BASE_URL = https://....railway.app

[zee.choudhry]

Hello @tlpui

Thank you for the additional information. It will be helpful if you Insert your workflow by clicking ‘</>’ and pasting in the workflow’s code

Check Logs

• Check http log and deploy log for n8n deployment for errors related to urls
• Check Railway logs for url related errors

Suggestions

• Clear browser cache. Reactivate the workflow and check the url in email sent or
• Use a different browser to reactivate the workflow and test it
• WEBHOOK_URL must match the callback URL in the Gmail credentials.

My Test

I tested local docker installation for n8n image and using similar workflow as your can not reproduce the problem with production url

[tlpui]

Hi,

Thank you for your replies. I tried to deploy n8n in another container, and after importing the flow from the previous one, everything went perfectly. It’s strange, but at least I have a solution now, and I’ll keep using the new one.

Thank you and best regards.