Nagios 2.x performs host checks serially, so it waits for the results of the individual check before checking other hosts in the direct vicinity to find out if the first host is DOWN or UNREACHABLE (see the figure in 4.1 Taking into Account the Network Topology). In contrast, Nagios 3.0 works in parallel here, provided that the parameter
max_check_attempts for the host to be checked is larger than
max_check_attempts=l the system behaves as in version 2.x.
In some cases this strongly influence performance, so you should always set a value larger than
1 in Nagios 3.0. Running host checks in parallel also means that a check result will sometimes get lost, since Nagios no longer waits until the result for every ...