Cloudron Forum

@hjball said in Kan.bn - The open source alternative to Trello:

Happy to assist with bringing Kan to Cloudron @mononym

Just let me know what you need from me.

Henry (author of the repo)

The biggest issue is that the deployment needs to happen in one docker container only.

For postgres, you use this: https://docs.cloudron.io/packaging/addons/#postgresql

For local storage, this:
https://docs.cloudron.io/packaging/addons/#localstorage

If you don't know how to use these add-ons, go here: https://docs.cloudron.io/packaging/manifest/

But essentially, if you can package it in one dockerfile only, creating the cloudron manifest is piece of cake.

Thank you!

It kind of does, but how about we'll add this to the logs to avoid confusion? The server already shows this, but the local command line doesn't.

I am a heavy user of the build service. Would it be possible to align the logging better between the server version and what I see in the console when running cloudron build? Specifically, it took me a while to figure out that the Build Service preferred a Dockerfile.cloudron instead of Dockerfile, which I assumed it would always use.

I believe the console logs don't provide me with that information (they only show the content of the dockerfile, but they would be similar at least, of course).
However, the app's logs on the server do specify this.

Also, I could see something like "you have multiple dockerfiles, which one would you like to use" or so.

@girish said in Install new custom app with environment variables?:

@ekevu123 cloudron install --env ENV1=1 ENV2=2 --location subdomain . Then, you can verify using cloudron env list.

Thank you! I think this is the best solution for this problem.

Yeah, I don't have the original, because I repaired this manually already, but that is basically it - the CNAME was a standard of Porkbun pointing to their server.

Well, it's solved quickly, albeit manually.

@girish Oh, really? How exactly do I use it? Do I need to upgrade the CLI?

When I install a new custom app using cloudron cli, I use cloudron build to build it and then cloudron install --location PATH --image IMAGE

However, for the installation process tu run through successfully, it requires to have environment variables set at that time.

I only know that I can set these variables using cloudron env set.

So, my current strategy is to run cloudron install, let it fail, terminate the process, set the environment variables, and then run cloudron update with the same image, which works.

I suppose the ideal solution would be to somehow use cloudron install but with environment variables set in the process.

Or should I simply set fallbacks and override them afterwards?

What's the best process for this?

Not sure if this is expected or a bug, but I noticed this:

1. I registered a new domain with porkbun
2. I connected the domain to cloudron and gave API access
3. I tried to install a new app on that domain using Cloudron CLI

Expected Behaviour:
Cloudron manages the DNS records automatically.

Actual behavior:
box:dns Attempt 4 failed. Will retry: Porkbun DNS error 400 {"status":"ERROR","message":"Create error: We were unable to create the DNS record."}

The solution was to delete the default CNAME entries manually that were pointing to porkbun (they were created when I bought the domain).

Technically, this was a conflict, and I assume cloudron, as it acts usually conservatively, would not interfere with the current state, but show an error instead until this is resolved? Or can cloudron only add and edit, but not delete existing entries?

I just wanted to bring this up, I am not sure if cloudron would be supposed to resolve this on its own.

Okay, thank you for the advise!

I am working on an app where users can create a profile page using a custom subdomain.

How would I handle routing? Do I have to use the cloudron API to add the subdomain to the app every single time? Do I set up a subdomain wildcard for my domain? Or is there any other solution for this?

You are right. You are forced to set up 2FA, then you log out and back in again, and only then the admin section appears under settings and you can enable/disable the apps.

Wow, I am impressed by who made that design decision.

@girish Maybe this could be added to the cloudron docs for this app as this is kind of important for using this, while it is far from obvious?

I have re-installed it just now, but I don't see what you see there. We can discuss it in the other thread.

I don't have Settings → Apps. In fact, there is no admin section after settings. The only Apps section is in the main menu, and if you haven't set these up beforehand, it will be empty. But to do that, you will need to have all your keys ready.

That seems to be group scheduling, but not really where you sync your calendar and send an appointment link to someone.

Interestingly, when I refresh the page, instead of agents, it quickly shows a model, and then switches back to "agents" with an empty dropdown.

Unfortunately, I couldn't resolve this with a few restarts.

Maybe I am not seeing it, but how do I start a conversation? I set up the api keys for openai, anthropic, and others, and restarted the app, there is a button "Agents" where I would choose a model to prompt, but the search is empty.

What caldav scheduling tools are you using to create an appointment link for others? I tried cal.com, but I dislike it actively because you can't change apps you use after the installation without re-installing, and I find Easy Appointments quite complicated, besides, I had sync errors with Radicale.

I haven't really found a nice solution that supports caldav, self-hosted or not.

I am not so sure if every app would need a specific MCP server written for it, because as I see it, the MCP server should expose only the cloudron-specific functionality of any app, not allow with the app itself.

What that means specifically: Using Claude Code, Cursor etc. I could prompt "check the production logs of my app to investigate why the hell this doesn't work".
Ideally, via MCP, the AI could execute test scripts that have been uploaded before via terminal, or do similar basic things such as go through files of the app.

I would be more interested in app-based interactions rather than the whole cloudron itself, but that use case might differ between people.

Using the Cloudron API is a good idea to solve this! That would enable MCP capabilities, I suppose. But I think the feature request still stands here, I understand I could code a workaround right now, but it could be convenient to activate the MCP where needed, perhaps by copy-pasting a snippet, which should be fine.