Re: Which frontend for SNMP monitoring of server farms?
On Tue, Oct 31, 2000 at 05:16:50PM +0100, Christian Hammers wrote:
> Which tools would you recommend for monitoring CPU/disk/squid-,apache-load
> of several servers at an ISP?
> Up to now I only have little watchdog scripts that bark via SMS/mail when
> some critical situation occurs but I want to check once a day on a web page
> and see some statistics and nice green signs saying me that everything
> was ok in the last few weeks and which server must be upgraded in the
> next few month or where a temporary max-load/bottleneck occured.
> tkined and openNMS seems to be suitable for this (maybe with "mon" as
> local watchdog) but I wonder if they're ready for production use.
> (At least openNMS looks very alpha)
I'm not sure why you'd use mon as a 'local watchdog'. Mon is quite
capable of monitoring remote machines. I've used mon for years and it
works great. It's a simple task to add new monitors for silly things
(silliest one I've done is to check if HP OpenView was still running...)
and monitor everything from status of PRI's to ssh running on remote
machines to whether the news spool is full...
With the 2 CGI's (monshow and mon.cgi) you get pretty web pages. If you
want graphs, use mrtg -- it's even possible to embed the mrtg graphs into
the monshow 'details' pages.
Between mon and mrtg, you pretty much have it all, and it's easy to
access from a web browser.