Mikael Fridh

~Is there an upgrade path for this? The `random_pet` name will require cluster destroy if upgrading the module.~

This probably doesn't meet your particular style and philosophies but it works. ~I'm going to restructure the top-level "flat" module list by calling it something else besides `modules`, perhaps `modules_set`~....

I just realized, and thus adding a note here. Exporting data which then also can cover this ordering need would be next level: https://www.runatlantis.io/docs/repo-level-atlantis-yaml.html#order-of-planning-applying Hadn't even thought about it yet...

Just making a note here related to my specific and current needs for why I wanted json output including module lists. - Atlantis seems to be getting functional module change...

Yes, would also like an option to exclude a resource entirely. In this case they are resources created by helm from an application that defines a helm chart via the...

Same. Causes issues in various transitions or mixes of `db.serverless` and other regular "provisioned" instances.

`refreshInterval: 0` meaning no updates even if the `spec` itself changes is a pretty wild breaking change. `refreshInterval` ... meaning "refresh" ought to mean polling or refreshing, on an interval,...

For now, "fixing" it with an override on top of the helm chart. Because the validation in the webhook is still useful for me on older Kubernetes versions. From [Tanka](https://tanka.dev)...

Seems to be backported to v1.0.2. See https://github.com/aws/karpenter-provider-aws/pull/6855, https://github.com/aws/karpenter-provider-aws/pull/6849 and then there are possibly several other open issues at least partially related to this issue and fix. Right?

It works but since the pricing stuff is done via AWS directly (not kube) - I need to explicitly ensure I have a (or preferably the matching) AWS profile selected...