Cloudron Forum

@nebulon given the significant issues that are cropping up with n8n breaking things with each release, how do we ensure we only pull the latest stable from GitHub which is where issues are raised and tracked? I've switched to manual updating because of breaking changes that come up from time to time.

@Joseph this is why I posted in support.

Hi team,

I've noticed that Cloudron keeps pulling pre-release versions of n8n?

Is there a setting somewhere to stop this and only pull latest versions?

@imc67 said in Per-application access rules:

@umnz said in Per-application access rules:

@imc67 I'm not sure what you mean, Cloudron does have a Trusted / Blacklisted IPs and Fail2Ban support.

https://docs.cloudron.io/security/

You are answering your own question: look at the subject of this future request and then the docs. What you mentioned is on server level .... not app level

Oops, my bad! Take my +1 instead lol.

@nebulon thanks for the insights.

Is it normal for Postgres to look like this all the time? It was like this at 256, 512, 1GB and now 4GB of RAM. The system information doesn't match the RAM allocation that's being reported.

@imc67 I'm not sure what you mean, Cloudron does have a Trusted / Blacklisted IPs and Fail2Ban support.

https://docs.cloudron.io/security/

What are your thoughts on removing the app proxy limit as a vehicle to help market/promote Cloudron and get people into the ecosystem?

@Package-Updates @girish I checked the n8n repository and 1.108.1 is pre-release? Any reason that got picked up by the package updater as stable/latest?

Hey everyone,

I'm currently running an instance of Cloudron with n8n (version 1.107.4) and I'm trying to determine what the best configuration is for performance out of the box since I'm experiencing frequent issues, laggy interface and occasional connection timeouts.

I have an instance with 4 vCPU + 8GB of RAM. By default, it looks like Cloudron allocates 256MB to Postgres and Redis.

When I run a n8n app, I cannot get it to exceed 100% CPU out of 400% available CPU. The UI frequently locks up when processes are being run - and upon further investigation it looks like Postgres is being maxed.

So I increased memory to 2GB for Postgres and Redis via the services panel in Cloudron. 2GB for Postgres was quickly filled and performance improved a little bit. But Redis use didn't change at all. I'm using whatever configuration n8n is set up for out of the box.

I then pumped Postgres up to 4GB and its currently sitting at 70% again, performance seems to have improved a little bit but I cannot get n8n to use more than 100% CPU of 400%. I've confirmed at the VPS level that only 25% of the CPU is being used. Does anyone know how to get n8n to use more CPU? It is set to 100% in the Resource Panel.

My workflows are relatively simple workflows that use scheduled triggers as well as web hooks. They basically check and/or send emails so the types of nodes I'm using are Outlook, Teams etc. Nothing I would consider to be insanely intensive. I'm also only running about 10 active workflows.

I'm hoping others might find this post and share their thoughts / configurations / tips / tricks etc.

Yeah I'd also be keen to know if anyone has got Redis working and what it looks like with a n8n instance. I don't know what's going on but n8n is getting slower and slower with each release (on Cloudron) and I can't figure out why. I think it has to do with queueing and/or workers.

I was recently helping a friend set up Cloudron and they were talking about how they liked the application proxy functionality after having dealt with Nginx Proxy Manager / Caddy / Traefik prior to Cloudron. It got me thinking about how to promote Cloudron to more people and have them dip their toes into the Cloudron ecosystem.

Do application proxies count towards the app limit on the free plan? If they don't, we should probably promote this aspect of Cloudron more since that makes Cloudron a very user friendly way to manage reverse domain name services -- especially in a home lab.

This may come out of left field, but are you running the app n8n on your Cloudron server?

@firmansi latest version was pushed to Cloudron

Yeah it's still pre-release unfortunately. So I can't update/test it yet.

Just an update for those who were affected and/or had workflows with advanced expressions - n8n have rolled back some changes and this release https://github.com/n8n-io/n8n/compare/n8n@1.107.1...n8n@1.107.2 which might have fixed the issue. Will have to wait and see.

No worries. I'd be screwed if it weren't for Cloudron's solid backup/update management. I've turned off Automatic Updates for n8n as a result of this - I recommend other Cloudron users do the same.

Confirming that there is a huge issue with n8n and I would encourage you all to check the status of your instances and resources as the failures are relatively silent if you do not have any type of error notifications built into your workflows. @girish

https://community.n8n.io/t/drastic-performance-drop-after-updating-to-1-105-2-large-workflow-execution-now-takes-60-seconds-instead-of-0-3/162566

https://community.n8n.io/t/performance-problems-with-the-1-105-x-and-1-106-0-releases/161353/19

Update: This has been fixed by the n8n team, and Cloudron release 3.97.0 uses the updated image. Things seem ok so far. I recommend turning automatic updates off for n8n as they seem to be moving fast and breaking things.

====

Has anyone else noticed that all releases of n8n on Cloudron after 3.93.1 are pinging resources and seeing frequent timeouts? One of the updates stopped you from being able to login at all because the port was mapped incorrectly.

Any n8n users here that use the Outlook or Microsoft nodes notice an increase in failed executions or timeouts?

I'm not sure if it is the Cloudron build but these timeouts are hurting a lot of processes and web hooks.

Is there a way to downgrade without restoring a backup?