Check_by_ssh benchmarks

[Karl DeBisschop]

On Thu, 2002-10-31 at 17:09, Sean Knox wrote:
> Currently we use check_by_ssh to check private resources (disk, uptime, 
> etc) on a small group of machines, as opposed to using nrpe. As ssh uses 
> a good deal more resources than nrpe, at what point would unreasonable 
> to use ssh to check machines? For the sake of conversation, say the 
> Nagios monitor is is a PIII 1gig with 256 megs of ram.

I seems like it's not often brought up, but when I wanted a lighter
alternative, I switched to check_snmp, since each nagios server I use
only operates within a fairly well firewalled environment.

I switched after about 100 services (10 each on 10 hosts). Not so much
because of the load per say, but when ssh fails, it does so much less
cleanly than snmp, in my experience.

Another reason is when a server is really hammered, say by a DOS attack,
ssh may fail, while snmp still works. I've had this happen, where it
took 15 minutes or more to get a session on ssh, but we were still able
to minitor the box. When we finally got in, we shut the port down
(typing blind because of the DOS) and everything cleared. But
snmp/netsaint never failed, and gave us the info we needed to quickly do
what was needed with ssh.

One other thing to keep in mind is that you can run multiple checks on
one check_by_ssh invocation. That could keep ssh viable longer.

--
Karl



-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en