[Traubenfuchs]

Deployment pipelines, CI/CD, (dev)ops. It's my biggest pet peeve ever that every company I have ever been to has written highly customised CI/CD/(dev)ops tooling that I had to fight with or adapt at some point. I don not understand why anything beyond heroku/dokku complexity is ever needed. The needs for teams to manage infrastructure is lunacy in this day and age.

[deckard1]

I really don't understand what has happened in the last few years in this space. At my company we went from stable but boring Jenkins to this Github/K8S hell. Everything is broke all the time. Literally every single day. It blows my mind. We are at the point that people have forgotten that things used to just work. It's like back in the day Windows98 users know you have to reboot the PC once in awhile. Because they know of nothing else. They don't know the past and they can't imagine a better, more stable future.

[Traubenfuchs]

I believe it‘s a conspiracy where sysadmins wanted to create their personal next generation job security.

Countless inane layers of overengineering no one can reason about anymore.

[nicolaslem]

Speaking of Dokku, I remember stumbling upon it a decade or so ago and dismissing it as a toy project that was unlikely to go anywhere.

Fast forward to the present, after noticing that the project is still going strong I did a deep dive in it and got really impressed. I think I will be moving my personal projects from the unmaintained Ansible mess I created to Dokku.

[Traubenfuchs]

With Dokku I can have a new dockerized webservice online behind https in <5 minutes. And I am confident I could set it up from scratch in <30 minutes.

All while not knowing a lot about devops at all.

[xakdog]

Yes! And some startups even set up k8s, manage db themselves, and spin up microservices. Even before having their first user!

So much hassle to support it when that shit breaks.

Also, to add to you list - Railway and Vercel are also super great