Site Does not respond after launching to Azure


I tried following this guide: Azure | n8n Docs
to test out self-hosting n8n.

After it launched, it tells me it’s running, but it’s not responding when I try to access it via IP or the Domain name I assigned to it.

Just using default settings for the service other than what was outlined in the guide. It is a
Standard_DS2_v2 node size.

I can successfully deploy the Azure Voting app to the cluster and access it via the public URL.

Any tips to get started or debugging information/recommendations?

Actually, have a little to add.
Looks like I’m seeing KubeEvents that have the errors:

  • 0/2 nodes are available: 1 Insufficient cpu, 1 node(s) had untolerated taint { }. preemption: 0/2 nodes are available: 1 No preemption victims found for incoming pod, 1 Preemption is not helpful for scheduling…
  • Back-off restarting failed container n8n in pod n8n-5ccf594478-r8r4s_n8n(071eb17f-1243-480d-a3c7-90b2cd039949)
  • invalid capacity 0 on image filesystem
  • MountVolume.SetUp failed for volume “ama-logs-secret” : object “kube-system”/“ama-logs-secret” not registered

Happy to share full logs via a PM.

Update here -
I got some help earlier today from one of my colleagues who works with Kubernetes. He found that we were getting errors about insignificant cpu, so we took down the cluster, and created a new one.

Now we are only having issues launching n8n container into the pod. Postgres and others are showing as running.

The note is: note: >-
Back-off restarting failed container n8n in pod

Does this help?

Hey @christianentole,

Welcome to the community :tada:

That doesn’t look like an n8n error, Are there any output in the logs from n8n itself? or can you share the full log output maybe?

1 Like

We did some more debugging. Looks like we had two errors.

  1. We had to patch one of the files, see screenshot below
  2. The password cannot be the same for both postgres users. We’re doing a test install and were just keeping it simple.

Seems to be working now!

1 Like

Looks like a good spot, I wonder why the password can’t be the same :thinking: The permission change is something that shouldn’t be needed and will likely be fixed in an update soon but it is handy to have a temporary solution thanks.

1 Like

You got it! We are up and running now.

Looking forward to testing this out internally.


This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.