We have a number of devices running Nagios located at various different sites, which are then supposed to report back to a single central site using NSCA…
Unfortunately, the volume of checks being performed at the individual sites means that a huge number of NSCA requests are coming back to the central box, and as the NSCA service seems to fork to handle each connection we end up with a massive load on the server.
Is there any way to improve efficiency, for example maintaining a single connection from each remote site which pipes the data instead of establishing a new connection for each notification?