Our plan is to have a monorepo with a bunch of stacks in it. Stack A is considered the "core" services like auth, DB, buckets, etc. Stacks B, C, and D are other things that depend on the core, like API and other microservices. Is there some way to set up

yarn start --stage devname

so that it only spins up a specific stack for them to work on, and will properly reference Stack A in a totally different stage?

Can you tell me more about why you'd like to have developers not get their own version of every part of the infrastructure?

it’s likely I just haven’t thought this through 100% yet.. 😉 but it seems like it’s just overkill. Cognito, Aurora, buckets.. there’d be duplicates of each of those for every developer. which also means that there would be a whole set of secrets & SSM params for every dev. not to mention, the first deploy would take 20 mins thanks to CloudFront distributions. but that would probably be a one-time thing it sounds like a hassle to maintain.

Got it. I've been opting to give every developer their own full setup especially, it will be slow on first deploy but it's a one time thing + easy clean up when you destroy the stack. Full environments helps them run tests that properly wipe + seed data without stepping on each other. Also you can make it uses Aurora serverless in local dev if you aren't using it in production which should help with the cost. You can disable certain resources if

scope.local

is true to speed up things as well That said if you don't feel like that's worth I have an idea for what you can do but haven't fully tested it. Basically you can override the stack name on only the non-shared stacks like so:

new SharedStack(app, "shared")
  new Stack(app, "individual", {
    stackName: `${app.stage}-${developer_name}`,
  })

This should make it so everyone deploys the same stack except for the individual stack. Set stage to

dev

or something and pass in

developer_name

maybe through an environment variable

interesting.. trying to wrap my head around it. what prevents

SharedStack

from getting deployed? what if I did something like this:

export default function main(app: <http://sst.App|sst.App>): void {
  // Set default runtime for all functions
  app.setDefaultFunctionProps({
    runtime: 'nodejs14.x',
    tracing: Tracing.DISABLED,
  });

  if (['dev', 'prod'].includes(app.stage)) {
    new CoreStack(app, 'core');
  }
  new APIStack(app, 'api');
}

so when a dev runs

yarn start --stage $(whoami)

, the core stack won’t get deployed

It won't be deployed but I imagine you'll want to pass values from it to the APIStack

ahh gotcha. I like that approach too

np!

I’m pretty sure I’ve decided to go with 100% isolation. so, each dev deploys their own stack.. and in fact each dev has their own AWS account too

That's what I do, aws account per dev

definitely

If you're using CRA https://docs.serverless-stack.com/constructs/ReactStaticSite#while-developing