DNS records are re-created even if heath check is failing

[trepel]

I scaled down my backend so that my DNS heath_check started failing. I manually removed the DNS records from cloud provider (AWS in my case). The records were re-created in AWS.

Unclear if this is the desired behavior or not, discussion needed.

[maleck13]

I don't consider this a blocker for v1 but I think we can document it. Really it is unadvisable to manually change DNS records managed by the controller.

There are other ways to achieve the same result:

• Delete the listener from the gateway
• If you have a DNSPolicy per listener, delete the dnspolicy for that listener
• (possibly) remove the DNSRecord directly (the health check should stop this from being re-published)

cc @maksymvavilov @mikenairn any other options for un-publishing a record set?

[trepel]

The third options works in rc10. The records are removed along with DNSRecord removal. The DNSRecord itself gets recreated but ends up with Not publishing unhealthy records so nothing is created in cloud provider. Everything gets sorted once backend is scaled up.

[maleck13]

nice this is what we will document so!

[maleck13]

Kuadrant/kuadrant-operator#1051