HAProxy Operator is a Kubernetes-native solution designed to automate the deployment, configuration, and management of HAProxy instances using Custom Resources to abstract the key components such as backends, frontends, and listens.
Is there a way to configure stats page to be exposed through a tcp port? If not, how do you handle stats in your setup? Do you somehow expose the unix socket outside the pod?
This question extends to exposing prometheus metrics (ie. http-request use-service prometheus-exporter in the frontend block)
It would be beneficial to add the check-sni backend option to the Backend resource, perhaps as a key in servers.*.check.
Motivation: without this option, the health check fails with Layer6 invalid response, info: "SSL handshake failure" error message. I noticed that it doesn't happen with every server that expects a TLS connection, but it bit me this one time.