I've been adding a bunch of checks via NRPE on remote nodes and this got me thinking. Isn't it inefficient to keep starting check_nrpe calls from the monitoring host all the time? <br><br>Why cannot nrpe on the remote node monitor some of the local services and only send a message back to nagios if there is a status change? <br>
<br>For warnings based on things like "disk usage", "cpu usage", "total procs", "pbs scheduler daemon status" , "cpu temperatures" etc. coudn't this approach relive the central host's cpu of a lot of endless check_nrpe calls? NRPE is already doing the work; its just a question of what initiates a communication channel between NRPE and nagios.<br>
<br>Just curious. Or maybe there is a way of achieving this already that I don't know of!<br><br>-- <br>Rahul<br>