According to the Red Hat documentation https://docs.openshift.com/container-platform/4.12/networking/ovn_kubernetes_network_provider/configuring-egress-ips-ovn.html, the maximum number of IP aliases per node is 10 - "Per node, the maximum number of IP aliases, both IPv4 and IPv6, is 10.".

Looking at the code base, the number of allowed IPs is calculated as
Capacity = defaultGCPPrivateIPCapacity (which is set to 10) + cloudPrivateIPsCount (that is number of available IPs from the range) - currentIPv4Usage (number of assigned v4 IPs) - currentIPv6Usage (number of assigned v6 IPs)
https://github.com/openshift/cloud-network-config-controller/blob/master/pkg/cloudprovider/gcp.go#L18-L22

Speaking to GCP, they support up to 100 alias IP ranges (not IPs) per vNIC.

Can Red Hat confirm
1) If there is a limitation of 10 from OCP and why?
2) If there isn't a limit, what is the maximum number of egress IPs that could be supported per node?

1.
2.
3.