Move managed cluster support (GKE) to general availability

[salasberryfin]

/kind feature

Describe the solution you'd like

As of today, support for provisioning GKE clusters is still an experimental feature and has been for a number of months.

In recent community meetings we discussed the viability of moving GKE support out of the experimental folder and into general availability. This would allow CAPG to follow other providers, such as CAPA and CAPZ which have full support for their corresponding managed clusters: CAPA, CAPZ.

The promotion of GKE to stable could be a major step towards a more feature rich and stable CAPG, and we could push to add support for existing GKE functionalities that are still not covered by the provider (there are a few related PRs).

Since this is an important change, we thought it would be great to create this issue where we can discuss pros/cons and plan what needs to be done if we decide to go ahead.

Initially, this seems to be a non-breaking change for the API, since it means adding types but not modifying existing ones, but it would be great to have feedback from those that are more knowledgeable of the codebase.

Anything else you would like to add:

As a follow up to this, ClusterClass support for managed clusters would be one of the relevant features we could be adding, as non-managed GCP clusters can already be provisioned from a class (refer to this) and this is expected to become a powerful/distinctive functionality of CAPI.

Please, feel free to share your thoughts on this.

I will be more than happy to take on/lead this if we decide to go ahead with the change.

[damdo]

@salasberryfin is looking for some feedback on this one from core maintainers
/assign @richardcase @cpanato

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[salasberryfin]

/remove-lifecycle stale

[salasberryfin]

Looking at unblocking this and moving the feature out of experimental.

Let's collect a list of PRs related to managed clusters that we should address:

• GCPManagedControlPlane changes to support various GKE features #1024
• Adding GCPMachinePool Functionality #901
• Add GKE NetworkPolicy support #1133
• Update node pool size even if autoscaling is enabled #1134

Feel free to contribute to the list with other relevant PRs/issues.

cc: @richardcase @damdo @cpanato

[richardcase]

/lifecycle frozen

[richardcase]

/area gke