New User? Register here - Existing Users: Username: Password: [Advanced Login]

 

 

Current Poll

Your preferred Interactive shell?









( 1376 votes ~ 15 comments )

 

Weblog entry #6 for Nilshar

See all weblog entries made by Nilshar.

XML logo

#6
System monitoring - collectd, munin, cacti, nagios
Posted by Nilshar on Mon 27 Apr 2009 at 08:50
Tags: none.
Hello,

I am looking for some knowledge about various system monitoring tools.
So far, I used cacti and nagios (both for differents things), and I've read a lot about munin (several articles/weblogs on this site), and collectd.

Soon I will have to choose one, and it will be a "no return" choice.

What I need is : rrd based tool, for local monitoring only, and it has to be easy to add/write plugins

During the few last years, I found cacti and nagios usefull, but not really flexible, and I don't have a good feeling about using it for what I need today.
Collectd and munin seems both very powerfull, easy to use, and easy to add custom scripts.. but I never used them..
And there is maybe (probably) some others tools I don't know yet.

So if you got some experience and knowledge to share, it will be welcome :)
[ 1 Comment | Add Comment | XML logo ]
<<< Previous Next >>>

 

Comments on this Entry

#1
Re: System monitoring - collectd, munin, cacti, nagios
Posted by daemon (146.231.xx.xx) on Mon 27 Apr 2009 at 21:01
[ Send Message | View Weblogs ]

I'm currently using both Nagios and Munin, as they do different things.

Going by your comment "it has to be easy to add/write plugins", you'll like munin. Simple Python, Bash, or Perl scripts can give you all you need, and the wiki is pretty good with regards to providing you with the info you need to expand on the example/included plugins.

Additionally, the two tools can be made to work nicely together (although I'm not doing that currently). For example, Munin can use Nagios' contacts, and can even be used to provide monitoring information for nagios.

There are two small issues with Munin that I don't really like, the rigid timing of the scans, and that currently you have to do one or more query/lookup for each graph. However, the latter at least, and possibly the former, will be sorted by the next minor-number/point release. The "multigraph" feature will be especially nice for monitoring networking hardware -- e.g. one bulk SNMP lookup will provide an individual graph for each switchport, rather than having to do a lookup for each port individually. And as for the scan timing issue, I think a more powerful CPU/more-RAM will clear up that issue for me too.

Cheers.
:wq

[ Parent | Reply to this comment ]

 

 

Flattr