Each message is preceded by a host or service check, which determines the current status. In the following two cases it generates a message:
One hard state changes to another hard state.
One computer or service remains in a hard error state. (The test therefore confirms a problem that already exists.)
To remind you: the
max_check_attempts parameter (see 2.3 Defining the Machines to Be Monitored, with host and 2.5 Defining Services to Be Monitored with service) defines in host and service objects how often a test should be repeated before Nagios categorizes a new status as "hard." If it is set to
1, this is immediately the case and is followed by the corresponding message. With a value greater than 1, the system repeats ...