Issue:

• Kubernetes API on OpenShift requires advanced health check configuration ensure graceful termiantion[1] works correctly. The health check protocol must be HTTPS testing the path /readyz.
  • Currently CAPA support standard configuration (health check probe timers) and for path, when using HTTP or HTTPS is selected.
• MCS listener/target must use HTTPS health check with custom probe periods too.

Steps to reproduce:

• Create CAPA cluster
• Check the Target Groups created to route traffic to API (6443), for both Load Balancers (int and ext)

Actual results:

• TCP health check w/ default probe config

Expected results:

• HTTPS health check evaluating /readyz path
• Check parameters: 2 checks in the interval of 10s, using 10s of timeout

References:

•  [1] API graceful termination: https://github.com/openshift/installer/blob/master/docs/dev/kube-apiserver-health-check.md 
• documentation: https://docs.openshift.com/container-platform/4.15/installing/installing_platform_agnostic/installing-platform-agnostic.html#installation-load-balancing-user-infra_installing-platform-agnostic 
• Current terraform setup: https://github.com/openshift/installer/blob/master/data/data/aws/cluster/vpc/master-elb.tf#L86-L94