Hello Jack,

I'm create a project with your template but this node is not working when using the node Form Trigger:

Cannot GET /form-test/xxxxxxxxxxxxxxxxxxxxxxxxxx

I've other problem when used the callback authentication Google API. I had use the URL primary.

Could you help me please?

Same issue, it seems the https://webhook-processor-xxxxx service doesn't work as expected….

Actually, it seems like webhook-processor only works in production. When testing you should use the primary domain.

Hi Jack,
Thanks so much for this templates! They are amazing.

Quick question:
What are the key performance differences between:
https://railway.app/template/5j7LO => N8N (w/ webhook processors) and https://railway.app/template/r2SNX => N8N (w/ workers)

I have a very specific use case where I need to do "Webhook" to "BigQuery" Insert and I need the fastest/most reliable performance possible. Which one would you recommend?

Also, based on this:
https://docs.n8n.io/hosting/scaling/performance-benchmarking/#example-multi-instance-performance

How much performance do you think I can get hosting this on Railway if I upgrade enough?
Will I get to 500/1000 requests per sec?

Happy to pay for a few hours of your time if possible.

re test link doesn't work

add the following variable to your primary node

N8N_EDITOR_BASE_URL

value: your primary node url

this means all test traffic go to the primary node

if anyone find out how to direct test traffic to webhook processor node, I would love to hear it.

Does anyone know where I can find instructions for setting this up on railway.app, which would help a rookie like myself get it working? I've imported n8n workflow jsons that were working on n8n.cloud, but I'm getting so many errors that it's unusable, including the 2 mentioned above.

Have you added N8N_EDITOR_BASE_URL to your primary node?

Have you added N8N_EDITOR_BASE_URL to your primary node?

How exactly do I do that, and will that allow the webhook node to work

Please see our documentation -

https://docs.railway.app/overview/the-basics#service-variables

re test link doesn't work

add the following variable to your primary node

N8N_EDITOR_BASE_URL

value: your primary node url

this means all test traffic go to the primary node

if anyone find out how to direct test traffic to webhook processor node, I would love to hear it.

This solution worked for solving the test_url for webhooks not working in n8n.

Thanks a lot @rckemal.

Originaly, in the primary, the variable "webhook_url" redirect the google drive (as an exemple) to the "webhook processor". The problem is that when, in n8n, I am trying to connect google "tools", the callback goes to the webhook processor and not the primary. Thus, I could not create my n8n credentials with google "tools". So, my solution was to change the "webhook_url" variable from my primary to the "primary url". Then it worked. My question is: Now, I suppose that my "webhook processor" is useless as my callback are all directed to my "primary url". Thus HOW do I configure correctly my primary variables to make sure my google callback, my webhook precessor and my primary works in harmony??

Thanks for bringing this to my attention. After your reply, I was able to get in contact with the n8n team to help resolve this and improve my templates.

You can fix this on your existing deployment by adding the following environment variable to the Primary service:

N8N_EDITOR_BASE_URL=${{RAILWAY_PUBLIC_DOMAIN}}

Thanks Jack!

Quick question, would it be possible to add a connection to the REDIS or Postgresql DB used in the template to store light workflow-related info? If REDIS doesn't support using the internal network, would Postgresql work?
I am asking because I tried connecting to the local Postgresql but wasn't successful yet.

Thanks!

After some attempts, I noticed that it allowed n8n to connect only when I disabled SSL in the n8n Postgresql credintials. Would this cause an issue considering it is internal network and should already be behind a firewall?

Thanks!

I wouldn’t use n8n’s database for any internal workflow use. It’s best practice to make a separate database using a template and connect to it internally so nothing gets messed up.

I’ve also never came across any problems as far as firewalls.

I'm stuck - workflow list just keep loading without actually bringing them up and also, can't seem to get webhooks to show https:// before the link - @jack - any chance you could help me troubleshoot? Happy to buy you a cup of coffee...https://www.facebook.com/yurashka/

That's odd, just deployed the template and couldn't reproduce this issue. Would you mind adding me to your project? jacklaffin@icloud.com

Thank you! Just invited you to both the N8N and the railway project

Thanks, just checked and added two variables for good practice that I just updated in the template, but not sure why you're getting that issue. Could you take a screenshot of the n8n workflow webhook element and the url it gives you?

Also noticed you had created two shared variables, but they won't do anything since they're not being shared to any services. Shouldn't need them anyways.

Thanks! any chance you can message me on facebook or maybe jump on zoom together?

1

I was able to resolve one of the issues by appending https:// to the beginning of the N8N_EDITOR_BASE_URL, my mistake.

However I wasn’t able to reproduce the loading state on the webbooks page. Is it easiest if I message you on Facebook?

Yes please, Facebook will be best! Were you able to access the Dropbox capture video I shared? Thank you again!!!

Yes, I was able to access it.

Perfect! Thank you! Talk to you on Facebook!

Hi,

I'm having trouble authenticating my Google drive account, made a few attempts starting from scratch but I keep getting this error message
"Cannot GET /rest/oauth2-credential/callback"

Any suggestion for me to try?

Thanks.

Hi Jack, any suggestions on what I could try?

Jack, I'm a newbie to n8n and Railway and installed your Queue Mode w/ Webhook processor template, which installed flawlessly. My question is does this template automatically update to the latest stable version of n8n upon redeployment, or do I need to manually perform a version-specific upgrade for each new release? Thanks in advance...

Hello,

I'm not jack, but all you need to do to update is redeploy the worker, webhook processor, and primary services, auto update for Docker images is not supported.

Thank you Brody, bailing me out once again. You make it sound very simple but please let me restate. All I need to do is redeploy worker, Redis (e.g., webhook processor), and then Primary and it will upgrade to the latest v1.77.3? I do not have to redeploy Postgres or update the Docker image itself? I ask as I had prompted Perplexity and it returned a short novel of step by step instructions that didn't exactly look easy to me.

Redeploy everything that is deployed from n8nio/n8n in the order I gave -

Perfect, thanks so much, Brody. Heading to do this now.

Hi Jack,

Thanks for the template!

I have an issue with the Telegram Trigger Node—it works fine in production mode, but in test mode, nothing happens, and I get a 404 error (screenshot attached).
The production webhook (/webhook/...) responds with 200 OK.

The test webhook (/webhook-test/...) returns 404 Not Found.

The N8N_EDITOR_BASE_URL seems to be set correctly. Could it be that /webhook-test/ should be forwarded to N8N_EDITOR_BASE_URL instead of N8N_WEBHOOK_URL?

Any idea what could be causing this?

Thanks!

Can you open your own thread, if we then decide changes are needed at the service level, we can bring them back here for Jack to change.

The deployment of the template is currently not working for me. Primary, Postgres and the Webhook processor deploy successfully. Redis does not start. Following that, Primary and Webhook both are crashing after numerous deployment retrys. Am I missing something or is this a general issue?

This would be because you are on the Trial plan, please upgrade to Hobby.

Thank you, that worked. I guess I did not see what parts of the trial plan were limiting me to make that deployment working. But fair enough, if I'm going to use the service even for trying out stuff, I can also pay the 5 bucks Thank you for the swift support, @brody.

Hi all! Got an email about railway infrastructure update - does anyone know if that would resolve ip4/ip6 issues that cause egress charges?

Hello,

No it won't, that is out of our control as it is caused by bad defaults with the ioredis package, what N8N uses to connect to Redis.

So until N8N exposes extra configurations for Redis, N8N will have to connect to Redis via the public network.

Hello!

I've been having an issue with custom nodes and wanted to know if anyone has had anything similar going on.

Everytime n8n releases a new update and I DON'T update right away, the custom nodes stop working. After I update n8n and reinstall the custom nodes, they start working again. Has anyone seen anything like it?

Obs: Not sure if I installed the nodes correctly. I just installed them using the UI and they worked fine.

I don't have the answer but I have the same issue.

For me it happens after I redeploy the services that I need to go in and delete + re-add community nodes.

del

Willthis fix this template as well, and what else is needed aside from setting this var to true? https://github.com/n8n-io/n8n/issues/13117#issuecomment-2642609099

New to railway and not familiar with how it works entirely, but is there a way to upgrade n8n to the pre-release version? They just released MCP and I want to have that on my server, but a simple redeploy only deploys with the latest release.

You can refer to my post on the workers version for making these changes: https://station.railway.com/templates/n8n-w-workers-6e5e2556#q8ys

As far was the webhook processor service, just make the same changes to it that you would for the worker service.

You can change the version tag for the docker image in the service settings to the pre-release version number.