Documentation is broken (fixed)

[moritz-br]

First, this is a truly wonderful project and I'm considering to use it for the production cluster of my side business application.

I've spent considerable time trying to locate any kind of instructions on how to run the docker container or generally start using this, finding absolutely nothing besides a few lines in the "Requirements" section that didn't help at all, until I realized that the documentation is in fact broken and the relevant parts are not present at all.

They are still indexed by Algolia and as such visible in the search, however (I was trying by myself to run the docker container and needed a token to login):

Clicking on this search result leads back to the index page of the documentation.

However, using the wayback machine, I finally found the instructions I needed:

https://web.archive.org/web/20241203133642/https://autobase.tech/docs#2-open-the-console-ui

I figure the relevant page(s) must've been deleted on accident. 🙏🏻

[vitabaks]

Thank you for your kind words and for sharing your experience!

Yes, the instructions for launching were intentionally removed as part of an experiment — we provided our email address instead to encourage users to reach out. The idea was to get direct feedback and offer personalized support for the first launch, including sharing the instructions and helping with setup if needed.

We’re always happy to assist, so please don’t hesitate to contact us if you have any questions or run into issues.

[email protected]

[vitabaks]

I’m still not entirely sure this was the right approach — I realize it might be frustrating for new users and potentially discourage them from trying Autobase. I’m now considering restoring the setup instructions publicly to make the onboarding experience smoother, while still encouraging users to reach out for support when needed.

[khawarizmus]

I really appreciate the honesty. I believe it's the right choice to make. When a product gains traction there will always be demand for paid support.

Also have you considered to offer a cloud version for autobase? Using hetzner as an underlying cloud provider could give you a competitive advantage in terms of pricing.

[vitabaks]

I’ve decided to bring back the Getting Started instructions to avoid any friction for users exploring this powerful tool. I absolutely agree that the initial experience should be smooth and accessible, especially for those evaluating Autobase for the first time.

As for a managed cloud version — that’s actually the opposite of what Autobase was created for. One of the core strengths of the project is that it can be embedded directly into your own infrastructure, giving you full control and avoiding dependency on cloud vendors. We believe this model provides greater flexibility, transparency, and long-term cost efficiency.

Regarding support, we feel that for production-grade deployments, having access to expert help should not be considered optional. With our affordable support plans, companies gain direct access to the creators of Autobase and PostgreSQL professionals. This not only ensures faster problem resolution but also unlocks expert guidance and peace of mind — a win for everyone. 🚀

[moritz-br]

As for a managed cloud version — that’s actually the opposite of what Autobase was created for. One of the core strengths of the project is that it can be embedded directly into your own infrastructure, giving you full control and avoiding dependency on cloud vendors. We believe this model provides greater flexibility, transparency, and long-term cost efficiency.

I think @khawarizmus was referring to a managed hosted control plane (autobase console) to deploy and manage your clusters (that would still be running on your own hardware / platform of your choice) - which is an excellent suggestion in my opinion. If I had found this offering I would've absolutely signed up and used it, even if it cost a few bucks.

In this scenario you could also offer your support plans and associated support functionality directly in this managed control plane, people don't have to bother running the docker containers and installing the console, and you could extend the control plane / console with more metrics, monitoring and other management functionality. It would be like GCP Cloud SQL or AWS RDS, but one to rule them all - completely vendor independent and applicable to any execution platform. I could see this quickly rising to extreme popularity.

On this note. I absolutely understand the need and wish to pay the bills. But I have to be honest, removing crucial documentation on a, to begin with, open source project and essentially forcing potential prospects to write an email is not a good choice and has left a bit of a weird aftertaste about this whole project in general.

I already considered using this in a production environment, was aware of your support plans and created an item in my backlog to sign up for a support plan with you once the application is live and cashflow has started. But now I've wasted literal hours trying to start testing it to begin with (by the way, it works fabulously well) and trust in the leadership of this project is a bit dented.

I will look beyond it and continue to explore your wonderful project, but if I may offer this recommendation from my personal experience doing so: More documentation would serve you more than less documentation. It's also taking me quite some time to understand, e.g., the specific configuration parameters I can configure in the command line vars, how to adjust the PostgreSQL configuration in the deployed machines themselves (e.g. how do I turn on huge pages?), and so forth.

That's nothing I personally want to write emails for, and nothing you should occupy yourself because your time is also more valuable than explaining these things over and over again to each interested person. 🙂