Slop alert. Use at your own risk
Selfhosted
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don't control.
Rules:
-
Be civil: we're here to support and learn from one another. Insults won't be tolerated. Flame wars are frowned upon.
-
No spam posting.
-
Posts have to be centered around self-hosting. There are other communities for discussing hardware or home computing. If it's not obvious why your post topic revolves around selfhosting, please include details to make it clear.
-
Don't duplicate the full text of your blog or github here. Just post the link for folks to click.
-
Submission headline should match the article title (don’t cherry-pick information from the title to fit your agenda).
-
No trolling.
-
No low-effort posts. This is subjective and will largely be determined by the community member reports.
Resources:
- selfh.st Newsletter and index of selfhosted software and apps
- awesome-selfhosted software
- awesome-sysadmin resources
- Self-Hosted Podcast from Jupiter Broadcasting
Any issues on the community? Report it using the report flag.
Questions? DM the mods!
Yeah, I was looking for that info, because I don't trust any (especially new projects, that use AI).
How did you know, it was AI though? I am just curious
The OP says so in the comments, but also:
- two week old github account, project created 2 weeks ago
- initial commit is a large 1.0 release with everything committed at once
- this also happens for 2.0 where everything is commited in one commit
- the code and comment structure
Also the look of the UI I can see default Claude UI a mile away these days. Always the same colours, fonts, layouts, ect...
That's right, I've been working on this project for a few weeks. I wasn't sure if I should commit it to a public repo, but I thought it'd nice to have other people testing it out, and giving their opinion. Honestly, I never used github before, that's why the account is new. I committed everything at once, when I felt like the application was "functional".
Git commits should be focused on one small thing at a time, doing small changes or additions. Then using descriptive commit messages like "add webook notifications" or "fix webhook triggering twice".
This makes the history of the project easier to see what changes were made and for what purpose. If something breaks webhook functionality then you can have an easier time finding what broke that feature. It also has the added benefit of someone showing their work progress. It's like you can be a math genius but if you don't show your work, your professor is going to think you cheated. In this case, it's a pretty dead giveaway that AI was involved. Since you are using AI then smaller commits also let you see what the AI to changing between iterations since even if you give it a copy of a file, it may change other parts of that file that you didn't ask for, and those changes could be good or bad.
And just because you are making these commits doesn't mean you have to "git push" them to your public repo right away. Since you're working alone and don't have any community contributors then you can just keep that work locally and push when you're ready for the next public release. Over time you could start using a development branch in git and push to that, so other developers can see your progress and when you're ready for a public release your merge the changes to your main branch. That's some more advanced git stuff but it's a core skill of any developer to be able to use branches.
Also there's nothing wrong with being new. I only mentioned the new account because when combined with the other things, this was a dead giveaway of AI use.
For what it's worth, good job putting together something that is useful for you. I would just encourage you to make sure you are trying to understand what you are doing and be willing to challenge yourself instead of using AI as a mental crutch like many people do. AI can only get you so far and you'll need to learn enough to call its bullshit and point out its mistakes.
Sorry, but you have posted only 1 sentence about the project and not even a link to the project.
Additional with the
scripts—basically "em dash" which is really popular among llm generated texts, i get a bad feeling about it.
there has to be a better way than giving a slop coded project access to your docker daemon (im assuming?)
Great idea. Automatic updates (e.g. Watchtower) make me a little nervous.
Automatic updates for bug fixes (e.g. 1.0.0 to 1.0.1) are usually fine - it's major and minor updates that are scarier. I've never used Watchtower so I'm not sure if it has an option to only allow bugfixes.
That would depend on each project properly using semver, which is unlikely.
Personally, I just risk all the updates. It's not a huge deal to recover.
For me, it’s all about finding the right balance. I don’t want to have to manually update for every little bug fix version bump. Most software I find that major.minor version tags, if they exist, are a good compromise with daily auto updates unless it’s a really fast releasing software where just a major version makes sense. I usually just track releases on GitHub or wherever the source is hosted and bump as I need. That takes care of probably 90-95% of the containers I run.
Who vigils the vigil?
That's a great question isn't it? That's why I posted it here, so maybe I can find people interested in working on this project and help me out to clean things up, get it more organized, structured and "free of AI slop". What do you think?
Can you provide a link to your repo?
Yeah absolutely, my bad. First time publishing things here and I thought it was attached to the post. https://github.com/kumucode/vigil.git
Copying my comment from the homelab community:
I haven't tried it yet, but here's some initial thoughts:
Does it support multiple separate docker-compose.yml files? It would be useful if it could pull the list of containers directly from Docker rather than having to paste the docker-compose.
Does it pull changelogs so that the user can tell if a change is a breaking change that'll require extra work?
It would be useful to support Webauthn/FIDO2 2FA instead of just TOTP. TOTP is being slowly phased out due to its weaknesses (it's phishable). Similarly, it'd be useful to support single sign on using OIDC (OpenID Connect) as a lot of self-hosters use Authentik, Authelia, or Keycloak to have one login for all their self hosted services.
Hi Dan, I'm also copying the answer from homelab community.
Thanks for your feedback. Much appreciated. For the first question, you click on add and past the image you’re currently using on your compose so the app creates a card with the current version. It’s a bit manual and tedious at first, but once it’s done, it’s easier to maintain. I think your idea is great to have the app just ¨find your docker-compose and do the work", but I don’t know how to do it yet. I wanted to test it manually first and see how it’d work out.
Vigil tells you if the newer version of the image is a major change or not. If you set it to update your compose automatically it will notify you and create a log, it something goes wrong you can easily revert it from the dashboard. Did I get your question right? Let me know if you meant something else.
Finally, security is an absolute must! I decided to use 2FA because most people won’t need to expose it to the web.They’ll probably use it on LAN. However, I do have adding OIDC (OpenID Connect) in mind, since many people indeed use Authentik, Authelia (these are the ones I’m familiar with). Since this is the early version, I didn’t want to make things too complex and also, I’m vibecoding it, so I’ll certainly need some experts out there to help me out to implement it correctly and safely.
If you have any question, just let me know and I’ll try my best to answer that.
Looks like a cool project. Starred.
I'm no tech expert either, so I'll keep an eye on how the community reacts to it, in terms of security.
Keep up the good work!
Make it support Podman next.
Their AI agent doesn't understand that.
I'd love too if I knew how to do that. I still have lots of things to learn and do before putting my toes on these waters, but I'm glad you showed some interest even to mention an integration with other tools. Thanks for that.
My god. This timeline is so horrible.
This looks fantastic! Great work.
Thanks man, appreciate it!