Right now the only health check mechanism for VRRP groups is a custom script. That's annoying for two reasons:
- The user has to write a script from scratch even for stock health checks like ICMP ping.
- The virtual server subsystem of keepalived does support multiple built-in health checks that we may want to add — but when we do, there will be a discrepancy with VRRP groups.
An ICMP check seems like a good start.
There's an additional reason to add ICMP check: currently every option of the old cluster subsystem maps to VRRP except set cluster group Foo monitor 192.0.2.1 (which monitors an arbitrary address using ICMP pings). Once there's a built-in command with equivalent functionality for VRRP, it will be possible to migrate old clustering configs to VRRP without any loss of functionality.