When working on a customer issue, we identified that there kube-apiserver were crashing. 

On further checking it was observed that even though the kube-apiserver was not healthy the lb was still sending the requests, leading it to crash.

In our documentation. the example lb configuration provides incorrect mode for checking master-apiserver availablity.

https://docs.openshift.com/container-platform/4.13/networking/load-balancing-openstack.html#nw-osp-configuring-external-load-balancer_load-balancing-openstack

Hence, please update it where it can check master-api url with /readyz for performing health check.

The load balancer must be configured to take a maximum of 30 seconds from the time the API server turns off the /readyz endpoint to the removal of the API server instance from the pool. Within the time frame after /readyz returns an error or becomes healthy, the endpoint must have been removed or added. Probing every 5 or 10 seconds, with two successful requests to become healthy and three to become unhealthy, are well-tested values.

eg : https://github.com/openshift/installer/blob/master/docs/user/metal/external_loadbalancer_ipi.md#haproxy-example-load-balancer-config


And if we can incorporate below KCS in our documentation as central place for checking all available health checks in loadbalancer.

1.
2.
3.