Migrate AWS to Simple Resource Class Selection

[negz]

Description of your changes

Fixes #41

This PR:

• Adds comment markers to make all providers, resources, and classes cluster scoped.
• Regenerates all generated code and manifests (using Support simple resource class selection crossplane/crossplane-tools#7) to support cluster scoped managed resources.
• Updates all controller logic to reflect that providers and managed resources are no longer namespaced, and thus they must specify which namespace their secrets live in.

It's a pretty huge diff but I've tried to keep everything broken down into logical commits to make the reviewing a little easier. Any changes to managed resource reconcilers that aren't mechanically removing namespaces and updating types are to ensure managed resource connection secrets are persisted to their desired namespace, and handle the case in which the (optional) writeConnectionSecretToRef field is nil.

I'm pretty confident in the claim defaulting and scheduling logic after testing it in per crossplane/crossplane-runtime#48 and crossplane-contrib/provider-azure#35. For this stack specifically so far I have tested:

• Creating a ReplicationGroup via a RedisCluster claim.
• Creating an RDSInstance via a PostgreSQLInstance claim.
• Creating an RDSInstance via a MySQLInstance claim.
• Creating an EKSCluster via a KubernetesCluster claim.
• Creating an S3Bucket via a Bucket claim.

At this stage I'm leaning toward not explicitly testing all the connectivity primitives. The changes to their types were very mechanical - they have no claims or connection secrets so it was just updating the logic that gets the provider secret. We should still test them holistically once all our in flight changes land - I imagine any bugs I might have introduced here would be quickly fixable.

Checklist

I have:

• Run make reviewable to ensure this PR is ready for review.
• Ensured this PR contains a neat, self documenting set of commits.
• Updated any relevant documentation, examples, or release notes.
• Updated the dependencies in app.yaml to include any new role permissions.

[negz]

@soorena776 fe6605e may be of interest. It seems the resource.CanReference argument to Build and GetStatus will be superfluous once all managed resources are cluster scoped.

[negz]

Uh, note to morning self: you forgot to add the new resource claim scheduling and defaulting controllers in this PR.

[upbound-bot]

[upbound-bot]

[soorena776]

@negz This looks great, thanks for putting all these together so quickly!
As for the passed CanReference object, they don't seem necessary here, but I think the api verbosity is still worth to support potential local referencers that we might come up later.
/LGTM

[upbound-bot]

[upbound-bot]

[negz]

Noticed #46 while testing this out.

[negz]

At this stage I'm leaning toward not explicitly testing all the connectivity primitives.

I actually ended up testing this using @soorena776's resource referencing work per soorena776/crossplane#1 and it has all worked.

[negz]

Looks like everything is working as expected. I'm going to hit merge.