feat: Support for wildcard domains

[jonstacks]

Closes #148

What

Allows for using wildcard domains via the ingress controller.

How

Domain CR naming

The algorithm for creating domains was to aggregate all of the rules.Hosts from ingresses and the listener.hostnames from Gateways. This would give us a list of ngrok Domain CRs to create/update. We currently replace . with - in the domain name to create the Domain CRs. This works fine expect that * are not valid in kubernetes resource names. So we now replace * with wildcard when generating the Domain CR Name. This makes the name now a valid k8s name.

This could cause future problems if you want to reserve the domain wildcard.example.com as the name wouldn't be able to be distinguished from the one generated by *.example.com as they would both generate a Domain CR named wildcard-example-com. We should probably look at generating these names like we do for HTTPSEdge CRs going forward to prevent this future problem, but this is not handled in this PR.

HTTPS Edge Labels

Previously we would add a label to the HTTPSEdge CR with the domain. This causes problems when using wildcards since * is not a valid character for labels. I initially debated about hashing the domain and using that as the label instead. However, we can extract the hostports from the existing edge. For HTTPSEdges generated by the driver, they will only ever have a single domain per HTTPSEdge and will always be listening on port 443. We can use this to compare if the edge in question is the same one as the one we are generating from the Ingress/Gateway configuration. This allows us to get rid of using the domain label or having to implement hashing for it.

Breaking Changes

No, with the caveat that if you could hit the above edge case temporarily if you are using both wildcard.example.com and *.example.com

[alex-bezek]

For the potential edge case of the conflicting * vs wildcard do we plan to address that right away, or is it something we should document somewhere for our users and support just incase someone hits it?

[jonstacks]

For the potential edge case of the conflicting * vs wildcard do we plan to address that right away, or is it something we should document somewhere for our users and support just incase someone hits it?

I think it is lower on the priority list. I'll write up an issue for it and link it so it appears in our list of issues.