Support node/pod affinity, tolerations and topologySpreadConstraints that CNPG supports

[nhudson]

The PR will add a new API field called affinity_configuration that is a type ClusterAffinity and topology_spread_constraints that is type ClusterTopologySpreadConstraints from CNPG. This will allow us to directly support all of the node/pod affinity and toleration types that CNPG supports.

The concept is that the control-plane will set the toleration/affinity when setting up the new instance, which we will just pass on to CNPG using their spec.

apiVersion: coredb.io/v1alpha1
kind: CoreDB
metadata:
  name: test-db-restore
spec:
  affinityConfiguration:
    podAntiAffinityType: preferred
    topologyKey: topology.kubernetes.io/zone
    nodeSelector: tembo.io/provisioner: spot
    tolerations:
      - key: "tembo.io/spot"
        operator: "Equal"
        value: "true"
        effect: "NoSchedule"
  topologySpreadConstraints:
    - maxSkew: <integer>
      minDomains: <integer> # optional
      topologyKey: <string>
      whenUnsatisfiable: <string>
      labelSelector: <object>

This change will move the default setting of affinity of CNPG from cnpg_cluster_from_cdb function into the CoreDB API with our current defaults

    podAntiAffinityType: preferred
    topologyKey: topology.kubernetes.io/zone

This will also do the same for the Pooler API and also the AppService Deployments.

The PlacementConfig does a lot of work specifically around converting the types. So we take in the ClusterAffinity and ClusterTopologySpreadConstraints they get converted into native Kubernetes API's and placed in PlacementConfig. The PlacementConfig has several methods that can convert the types to use with the Pooler or we can use the types directly in a normal Kubernetes Pod or Deployment if we wish.

fixes: PLAT-431

[ianstanton]

LGTM once everything is passing 🎉