This message was deleted.
# general
s
This message was deleted.
l
Dependencies between stacks in the same project sounds like a very bad idea to me... how do you deploy stack A in an environment that doesn't have stack B?
You need the
--suppress-outputs
option for this, btw
a
So Stack A controls DNS. Stack B is just another env. I need to export some AWS ACM DNS records for automatic cert verification….
l
I recommend moving DNS to another project, that must be deployed before any stacks in this project.
Have all stacks in the current project be dependent on the only stack in the DNS project.
a
Does pulumi support inter-project dependencies?
l
Absolutely.
It is core to most implementations. It would be common to have projects for base infra + users/auth (deployed almost never); shared infra like Config, Orgs and DNS (deployed only when needed); shared app infra like Cognito, buckets and databases (deployed on appropriate feature release); apps (deployed on demand or feature release).
a
Ok — this sounds like exactly what I need!
Do you happen to know off hand a good starting place in the docs for me to look at?
l
Generally, group into projects based on deployment cycles; group into stacks based on location (meaning anything like region, staging vs. prod, dev1 vs. dev2, etc.).
There is a page of recommendations about this, iirc... onesec...
a
Thank you!
👍 1