This message was deleted.
# general
s
This message was deleted.
f
guess stacks need to be in same project :S https://github.com/pulumi/pulumi/issues/3395
a
The main problem is that we can't use multiple backends, e.g., both the Pulumi service and S3. Is that your case too?
f
I just had different projects in different storage paths, e.g. azblob://pulumi/projecta and azblob://pulumi/projectb ... but couldn't get that wroking
a
I've just used my own naming convention on top of stack names. So a stack name wouldn't be simply
dev
or
stg
, but
project-name.dev
,
project-name.stg
, etc. And I store them all in a single Azure storage container. This affords me cross-project stack references.
Something along those lines is described in this blog post: https://blog.scottlogic.com/2020/04/21/starting-with-pulumi.html
r
You can create an azure key vault as part of the pulumi program and store the values in there. Then in other stacks you can azure.keyvault.getSecret()
if you go this route, be sure to grant your current identity read/update/delete access policies to retain control of the key vault 🙂
f
I went with one project with multiple stacks .. seems to work okay so far
a
That's fine. You'll have to split into multiple projects just when/if you need to deploy parts of your code separately. With a single project it's all or nothing — all infra code is executed.
f
yeah I'm looking to bring more devops people into this, so would prefer to have things properly split up instead of all or nothing. Some of these things are also quite unrealted