System & processes monitoring recomendations?

[Edwin Groothuis]

Hello Anton,

On Sun, Jan 21, 2007 at 06:00:40PM +0200, Anton Blajev - Valqk wrote:
> I'm wondering what tool to use for system and processes monitoring?

We use Nagios in combination with the Nagios Remote Process Executor
(NRPE) and a handfull of homewritten scripts to:

- Check remote accessibility of services (that's formal talk for
  "setting up a TCP session and checking the output")

- Checking the local availablility of services (that's formal talk
  for "check if the process defined in the /var/run/foo.pid does
  exist") on remote machines.

- Check for the consistency of systems (that's formal talk for "I
  expect these processes to be running, and these network cards to
  have these media features, and do we have the required BGP
  neighbours, and these not more than X users logged in, are the
  mailqueues not too full, and etc")

For the 139 devices (FreeBSD machines, Linux machines, switches,
routers, radio links, Windows machines and even serial-to-ethernet
convertors), we monitor 1146 different services. And despite the
huge amount of things, I get on average only 10 messages a day,
mostly from one of the SMTP servers when a neighbouring body is
sending out their daily spam-run to all of their recipients (incoming
mailqueue, CPU load, mail spike :-)

Edwin

-- 
Edwin Groothuis      |            Personal website: http://www.mavetju.org
edwin at mavetju.org    |          Weblog: http://weblog.barnet.com.au/edwin/