Host reported down while local checks are successful and ping is ok

For multiple hosts I configured hostalive checks via

object Host “host.fqdn” {
import “generic-host”

address = “host.fqdn”
vars.agent_endpoint = name

vars.os = “Linux”
vars.type += [“http”, “ssh”]
}

template Host “generic-host” {
max_check_attempts = 3
check_interval = 1m
retry_interval = 30s

check_command = “hostalive”
}

However, a single host is flapping several times a day between up and down. Strangely a ping command running on the master does not report any package loss during such a time. All local checks for that host are reported “OK” continuously. Both, the log file on the master as well of the agent did not contain anything useful to me to debug the problem.
Any suggestions on why these checks are failing?