feat: Make auto_provisioning_defaults a non-beta feature

[stanley98yu]

Fixes #1008

This PR simply makes the auto-provisioning defaults block not beta-exclusive, but I also want to add the min_cpu_platform argument in the block as specified here since it seems we added the feature for default node pools in #1057. However, I don't understand why we pull the value from the first node pool using lookup(var.node_pools[0],...) because, from what I understand, min_cpu_platform is a cluster-level variable, so it feels weird to read from a specific node pool's arguments, like pool-01, even though it will apply to other node pools as well. Shouldn't it be its own variable and read in like var.min_cpu_platform or something similar?

[morgante]

Shouldn't it be its own variable and read in like var.min_cpu_platform or something similar?

It's actually not a cluster-level variable. min_cpu_platform is part of the node pool config for particular nodes.

The cluster-level setting is for node pool autoprovisioning. ie. what setting to use for any autoprovisioned node pools. We've made the decision to have your first node pool be your "default" node pool and the settings from it are mirrored to any autoprovisioned node pools.

[comment-bot-dev]

Thanks for the PR! 🚀
✅ Lint checks have passed.

[stanley98yu]

That makes sense. Thanks for the explanation! The min_cpu_platform for auto_provisioning_defaults mirror the first node pool's value as expected then.