There is now a blank public repo for the "fully-featured, infinitely scalable IaaS with predictable pricing." It's pretty minimal, but it's good to slow down and think about how things are laid out. The further I get into this project, the more I question whether a fork-able repository or a CLI that generates boilerplate is the right approach. Either way, it will be valuable to prove that all of these dependencies work together.
Repository: https://github.com/winstonpuckett/sunflower
I've also narrowed what the experience should feel like. A user who has never heard of Sunflower should be able to complete "hello, world" within 5 minutes. This might be hard given the prerequisites (like, "have a Kubernetes cluster already").
Top comments (3)
I had been considering that route - a cli would also allow generating a service map which could be cool. I've of the hardest things I've found at scale is just knowing exactly what is being run in prod and what services each service depends on. Thanks for sharing!
Honestly, this project would be a lot of work and I'm thinking I might not follow through now that I'm seeing it might not get a lot of traction.