Hey all, We’re coming from the dotnet world, and ...
# general
a
Hey all, We’re coming from the dotnet world, and we’re exploring Pulumi as our next IaC solution, comparing it to other leading IaC solutions We think that we have an overall understanding of how the product works, but we still lack understanding of a couple of things: • We understand that it’s a state-based solution, which hold information about the deployed stacks, which can exchange information using `StackReference`(s) and outputs. Having a single state-backend is potentially a huge SPOF (single point of failure) for us (please feel free to change my mind, I’m open). Is there a way to reference other stacks which are deployed in other state backends? • A stacks full “id” ? appears to be
{organization-name}/{namespace}/{stack-name}
(please feel free to correct me). This doc claims that the
{organization-name}
is something you can configure, however locally it always seem to default to
organization
? Is this something that’s available only in Pulumi Cloud? I’d appreciate your input 🙂
s
WRT to your second point: when using a DIY backend (not Pulumi Cloud), the organization name is indeed always
organization
. I believe we have an open issue tracking the ability to configure/change that. As for cross-backend stack references…I’m not aware of any such functionality planned on the roadmap.
a
Thanks @salmon-account-74572 . That's interesting, for instance in Terraform this seems to be common functionality https://developer.hashicorp.com/terraform/language/state/remote-state-data
s
I understand. I’m not sure why this sort of functionality hasn’t been implemented yet.
a
I'm wondering if TF state is equivalent to 1 stack's state, I'll ask around
@salmon-account-74572 another set of questions that just popped up: • https://www.pulumi.com/docs/concepts/update-plans/ this appears to have been in preview for the past 2 years? Any news on this feature? Is this headed to GA or to the scrap heap? • We've done some internal proof-of-concepts, one of them was about deploying resources to Kubernetes using HELM. We've noticed that private OCI registries appear to be problematic. Is this something that's on the roadmap?
s
Let me make some inquiries internally so I can get the latest info on anticipated future updates.
a
thank you! looking forward to what you have to say
s
Regarding private OCI registries: our engineering team confirmed that it's a highly upvoted feature and it's near the top of their list, but I don't have any committed timeframes as yet: https://github.com/pulumi/pulumi-kubernetes/issues/1914
a
and what about the update plans?
Another interesting thing we’ve noticed: when managing helm charts, we can only provide value overrides via
IDictionary<string, object>
. Are there plans for Pulumi to manage multiple HELM values files?
s
I don't have any additional information on update plans yet. As for multiple Helm values files, your best bet is to check https://github.com/pulumi/pulumi-kubernetes/issues to see if we have a tracking issue there. If not, I would suggest creating one; if there is one, then upvote it.
a
will do
thank you @salmon-account-74572. much appreciated
s
NP, happy to help!
s
@salmon-account-74572 is there a possibility to get some more information about "update plans"? It seems to be such an important feature, which is crucial for our usage, and we wonder "how mature" it is, and when we can expect it to be in GA. Also, being 2 years in the experiment phase seems to be "suspicious".
s
The only thing I can point you to is our public roadmap: https://github.com/orgs/pulumi/projects/44 I don’t have any additional information beyond that.