Make Cluster API cluster creation repeatable

[puja108]

User Story

• As a Giant Swarm user I want a repeatable way of creating Workload clusters with Cluster API (CAPI) so that I can achieve immutability of my infrastructure and rely on a specific version in my testing.
• As a Giant Swarm engineer I want a repeatable way of creating clusters with CAPI so that I can create automation around releases for testing use cases.

Details, Background

Giant Swarm currently versions all components and changes to Workload Clusters within a Giant Swarm release version that customers and internal teams rely on in their testing. Customers often test a version before they upgrade production clusters to that version or have a gradual rollout approach of versions from environment to environment (dev to staging to prod). We aim to achieve a similar level of repeatability with our adoption of upstream CAPI controllers.

[puja108]

By now we have a repeatable way to start CAPA clusters on AWS.

We are deploying the controllers in a controlled manner using our existing release concept (codename v20) and creating the clusters with our kubectl plugin kubectl gs.

Azure is also aligning with this way of cluster creation, which will enable us to move on to #298

[puja108]

with the cluster apps we have this, closing