If you're solo, you probably know enough that fix-forward is the way to go. (This assumes the system isn't critical enough that it needs fixing even if you're asleep, but then you would probably not be solo anyway.)
1: Work on something that you yourself or your users want to exist. Don't think about what others might want. Make something that you consider pretty cool shit. Or something your users are asking for. Something where you or your users say "OMG when this is ready I will fricking love using it! I can't wait!".
2: Most founders, me included, often get stifled. They stop putting time into their project because they think it is just a dream. Remember that all amazing things have been built by people just like you. And to keep putting time into building something that you yourself or your users want.
Not immediately, but in the following weeks.
Thanks again for sharing.
PD: I follow a similar approach. I am glad you said it.
If I want to set it up on a new server, I start a fresh Debian machine, do "apt install mariadb-server" and I am good to go.
I don't like managed services (like Heroku) because:
1: I cannot easily replicate the setup locally.
2: It creates vendor lock-in which will cause me trouble when the vendors service changes in a way I don't like.
3: I would not be able to put the whole setup of my project in one setup script. It is so nice to just have a single script that contains all the steps that need to be done (Like the "apt install mariadb-server") to run the project on a fresh machine.
It's easier than you think if you stick to your distributions' packages. Just need to know how to 1, do auth right 2, enable SSL 3, run with SELinux 4, set a different location (for attached storage) 5, managed backups.
Here is what works for me:
CI: From the terminal, I run my tests and commit to git.
Deployments: rsync
Rollbacks: Never did one. If something breaks, I fix it and rsync the fix to production.
DB: MariaDB
k8s: I don't use it. Computers are very fast these days. A cheap single VPS will get you a long way.
Nightmare: Not really. I spend about 30 minutes per week on DevOps.