If you test services with local plugins (see Chapter 7) via NRPE (see Chapter 10), all these tests will come to nothing the moment the Plugin Executor fails. With service dependencies you can prevent Nagios from flooding the appropriate administrator with messages on the dependent services. Instead of this, the system informs him specifically of the NRPE failure.
Aa with such service dependencies, Nagios also has host dependencies, which suppress messages, depending on individual hosts. Both variations can also be used to specifically "switch off" tests.
Let us take as an example the host
linux01, illustrated in Figure 12-4, on which locally ...