Eg, do you all just create separate apps for each layer? How do you typically manage auto-deployments?
We have separate apps for “staging” vs “production”, usually. I know some people have used our API to automate branch apps, but there’s no first class support. Databases make this hard and we haven’t quite decided how best to handle them yet!
Do you know any examples of ppl having created PR / branch apps? We are moving to Fly but our devs and PMs love our PR instances our current provider has as a first class feature.
You are right, the DB makes it tricky but we’ve seen that a staging app whose DB can be considered somewhat ephemeral (dropped / recreated / seeded) helped streamline it.
Let me see if I can find any! Are you using something in particular for CI? GitHub actions maybe?
We are using Render’s Pull Request Previews with GitHub right now. But something with Github Actions would work well!
Edit: Sorry, yes we are using GitHub for CI right now too.
Usually when we want staging/production, we want to make sure we can prevent to make mistake for staging/production. Noticed that the fly can create different org, so for me, I will create two org, one for production, one for staging. But for billing, if fly can have ability that we can create diff envs under same org, that would be really useful.
Would shared billing across organizations do what you want? I’m a fan of that model, and I think it’s what we might end up doing.
yeah, that would really useful.