Remove dependency to Active Cluster - active cluster name

[nilsgstrabo]

External DNS, app alias and active cluster ingresses are only created when cluster name matches active cluster name (defined in flux).
This made sense when we used external-dns to managed DNS records in the DNS zones based on ingress objects in Kubernetes, but now that we use wildcard DNS records there is no need to differentiate between active and non-active clusters when it comes to ingresses. All ingresses should be created regardless of active cluster state.

• Cleanup in radix-platform - https://github.com/equinor/radix-platform/blob/master/scripts/move_custom_ingresses.sh
  • Cleanup migrate custom dns / aka. active cluster from platform radix-platform#1478
• Cleanup in radix-operator
  • Create all ingresses always, independent of active cluster #1213
  • Remove usage of active cluster name #1217
• Fix in radix-flux

DoD
Works as before, operator will not start creating/removing ingresses when a new cluster is deployed

[Richard87]

Currently RADIX_PUBLIC_DOMAIN_NAME is reverted to contain the domain from RADIX_CANONICAL_DOMAIN_NAME when its not active.

• this can help with testing during cluster migration (if the application is using RADIX_PUBLIC_DOMAIN_NAME during auth for example to redirect back to the application, and they are using their own oauth provider)

• but it can also be wrong during migration (if they are submitting the domain to 3rd parties for callbacks, if the cluster is shut down before the callback, the will lose the information, if they assumed RADIX_PUBLIC_DOMAIN_NAME always points to the active cluster/application)

I would like to simplify the logic, this is also the last step for the operator to not need to know what cluster is active

We have decided to simplify, RADIX_PUBLIC_DOMAIN_NAME will always be public, and never changed to canonical in standby-clusters