[sclug] Monitoring and Alerting

[Sapan Ganguly]

Dudes and Dudettes,

I have probably asked this before but can anyone recommend some good
software for monitoring and alerting on a network of Linux machines? 
There are a few Solaris machines too, it would be nice if I could
monitor them also but I'm not too concerned about them since they will
be retired as soon as we get around to it.

I'm looking for something that is extremely easy to set up and does
not require extra agents or daemons to be installed on the machines. 
I would prefer to use SNMP, I would like to monitor memory usage, disk
usage, CPU, IO and if possible the state of the RAID arrays.  I would
also like to monitor specific daemons/programs to make sure they are
running.  Checking that certain ports are available and that they give
the correct response would be good too, SMTP, HTTP etc.

When things go wrong, are not available or reach a certain threshold I
would like an email to  be sent to me.

A couple of people here have spent about 3 months trying to make HP
Openview do something useful and have failed.  I've looked at Nagios
but it looked very fiddly to set up.  Cheops looks like it will do
everything I need but it is no longer developed therefore there are no
patches to stop it crashing.  Cheops-ng does not seem to being
developed anymore either, it has no alerting facilities and is
therefore useless for me.  NINO looks pretty good but I am still
trying to make it do everything I want, the alerting won't work for me
yet.

Any suggestions?

Thanks,

Sapan