vic on Nostr: binkle mangeurdenuage :gnu: :trisquel: :gondola_head: 🌿 :abeshinzo: :ignutius: ...
binkle (nprofile…m5h7) mangeurdenuage :gnu: :trisquel: :gondola_head: 🌿 :abeshinzo: :ignutius: :descartes: :stargate: (nprofile…y2fr) Oh, well first you need to make sure you pick a framework (for both backend and frontend) and then define your domain models, which you will ensure are valid with JSON schema. Then pick a database and database driver, plus migration and ORM libraries.
Then you have to make sure you get a consistent dev environment set up, so get your Prettier config written and debugged, and write a Dockerfile with stages for both building and running. You'll want to make sure that's set up with a CI system (there are only about 48 to choose from), and set up Renovate to keep your dependencies up to date. You also need to make sure the images are constantly scanned to keep the vulnerabilities out, since you'll be depending on approximately 10,000 different JavaScript libraries.
Once you have in containerized, you're good to go! Hah, just kidding. Next you need to make Kubernetes manifests for pods, services, storage, and ingress. You can try to look up the "right" way to do it, but it will be deprecated by next week, so just hack away until it works.
Then you have to package all those manifests together as a Helm chart, which must be independently version controlled and managed, with its own tags and release cycle. Oh, and you have to make sure it's super well-documented, and ensure you include every possible field that someone could ever want, or you'll be hearing about it in your Issues page later.
But wait, there's more! You can't just deploy that Helm chart, you have to do it as code! So it's time to learn Flux for continuous delivery. Learn about datasource configs and the Helm controller! Set up webhooks between the Kubernetes cluster and your Git repo! Pick how often to sync! Decide whether CI should be allowed to trigger Flux on main or whether you want to handle it yourself!
Oh, but we're just getting started. You still have to figure out ingress, certificates, high availability, storage, backups (and practice restoring!) and more!
Welcome to the hell that is the modern web!
Then you have to make sure you get a consistent dev environment set up, so get your Prettier config written and debugged, and write a Dockerfile with stages for both building and running. You'll want to make sure that's set up with a CI system (there are only about 48 to choose from), and set up Renovate to keep your dependencies up to date. You also need to make sure the images are constantly scanned to keep the vulnerabilities out, since you'll be depending on approximately 10,000 different JavaScript libraries.
Once you have in containerized, you're good to go! Hah, just kidding. Next you need to make Kubernetes manifests for pods, services, storage, and ingress. You can try to look up the "right" way to do it, but it will be deprecated by next week, so just hack away until it works.
Then you have to package all those manifests together as a Helm chart, which must be independently version controlled and managed, with its own tags and release cycle. Oh, and you have to make sure it's super well-documented, and ensure you include every possible field that someone could ever want, or you'll be hearing about it in your Issues page later.
But wait, there's more! You can't just deploy that Helm chart, you have to do it as code! So it's time to learn Flux for continuous delivery. Learn about datasource configs and the Helm controller! Set up webhooks between the Kubernetes cluster and your Git repo! Pick how often to sync! Decide whether CI should be allowed to trigger Flux on main or whether you want to handle it yourself!
Oh, but we're just getting started. You still have to figure out ingress, certificates, high availability, storage, backups (and practice restoring!) and more!
Welcome to the hell that is the modern web!