I’ve configured n8n to run on ecs, the main process and the worker process seem to come up fine but when try to access the webhook instance I am seeing errors. They are hosted on separate dns entries, so it will be n8n.company.comn8n-queue.company.com and n8n-webhook.company.com, so routing of traffic does not seem to be an issue.
Hi thanks for getting back to me. No I haven’t actually so I will try that.
One thing as well that I should have mentioned is that invite links do not work either, the invite link is generated as https://n8n-webhook.company.com/signup?inviterId=blahblahblah and this returns Cannot GET /signup. Is this something I have misconfigured?
Hi @fclark182 I suspect this could indeed be a problem with the configuration. Your webhook processor wouldn’t need a public hostname on its own (though I can see how our documentation suggests this could be a possible approach, I’ll suggest a change for this internally).
Instead you’d have a single public hostname pointing to your load balancer (or reverse proxy), then distribute requests as described here:
Redirect any request that matches /webhook/* to the webhook servers pool
All other paths (the n8n internal API, the static files for the editor, etc.) should get routed to the main process
Perhaps you want to give this a go?
Alternatively you can also replace the webhook hostname in https://n8n-webhook.company.com/signup?inviterId=blahblahblah manually with your main instance hostname for a quick test, but this is of course not a long term solution.