The Nagios thread

Discussion in 'Business & Enterprise Computing' started by Mac, Jun 30, 2009.

  1. Mac

    Mac Member

    Joined:
    Aug 1, 2001
    Messages:
    762
    So I finally bit the bullet and jumped onboard the Nagios rollercoaster last week.

    I've always been a Windows Admin and had very, very little to do with Linux. As a 'Network Admin' who deals primarily with Education customers, they all use and have extroadinaly good Microsoft Licencing - so much so, that the cost of the software for my clients is taken out of the equation when they look at Windows Vs Linux. Anyway - this isn't a software windows vs linux thread...

    So now I'm learning Nagios. And Ubuntu. And I'm confused. :)

    I have my installation up now though, and its monitoring 14 servers quite nicely with my Email Alerts working too!

    First impressions - I'm finding Nagios (and Ubuntu) well documented. The offical doco is clear and concise however i find on occasions they assume I know more than I actually do. Overall - I like it though.

    Nagios though, to use its full potential appears to need a few 3rd party plugins. These are troublesome - I find the documentation often assumes I know alot more than I actually do. And often I find what doco that their is, confusing and sometimes contradictory, or in a foreign language...

    Who else is using Nagios? Recommend any plugins?
     
  2. QuakeDude

    QuakeDude ooooh weeee ooooh

    Joined:
    Aug 4, 2004
    Messages:
    8,435
    Location:
    Melbourne
    We've started using it, and its great - one of our guys is a Linux guru, so there was no issue whatsoever in setting it up. We're actually running it in parallel with IBM's official monitoring of our servers, and it hasn't missed a beat - I'll typically get a Nagios alert before I get the IBM one, so its certainly proving to be a stable and reliable tool. We've tested our external SMS module with it too, and that works a treat. If we decide to use this as our main production monitoring system then we'll go live with that part as well.
     
    Last edited: Jun 30, 2009
  3. Gecko

    Gecko Member

    Joined:
    Jul 3, 2004
    Messages:
    2,715
    Location:
    Sydney
    I use it every day to monitor a worldwide network with more servers than I care to count.

    I have it up on a secondary monitor on my desk - screen goes red the moment that any issues are detected. I can normally get an alert and have solved the issue before anyone else notices theres a problem.

    Yes, it isn't the easiest piece of software to set up in the first place, but once its up you'll wonder how you survived without it.

    Cacti is another good tool as well, gives you some nice graphs so you can monitor trends and resource utilisation over time.
     
  4. OP
    OP
    Mac

    Mac Member

    Joined:
    Aug 1, 2001
    Messages:
    762
    Was looking at Cacti but found it really confusing to set it up... I ended up using http://www.pnp4nagios.org/pnp/start - which I still found confusing but managed to get working...
     
  5. Spingo

    Spingo Member

    Joined:
    Jun 28, 2001
    Messages:
    1,052
    My biggest problem with a lot of the plugins that are out there is that they don't do what I need them to do. A classic example is the check_disk plugin that comes with the nagios_plugins project. This plugin allows you to check how much disk space is free on a logical disk - but to check free space as a percentage, you HAVE to use an integer value (see http://nagiosplugins.org/man/check_disk). That's useless when you're dealing with 2TB LUNs.

    Another similar example is the check_disk plugin that is bundled under NSClient++, where in addition to the constraint mentioned above, it also doesn't support LUNs mounted under a path.

    So more often that not, I develop my own plugin scripts. I have a fairly messy collection of scripts that expand upon what other scripts do, that monitor other things that I like to monitor (including a bunch of scripts that monitor HP Server hardware in detail) as well as the scripts that check the proprietary server-side applications that we develop in-house.

    I think Nagios' best feature is that it can be expanded in a way. As a result, I no longer call Nagios a systems monitoring application. It's more of a framework.
     
  6. username_taken

    username_taken Member

    Joined:
    Oct 19, 2004
    Messages:
    1,352
    Location:
    Austin, TX
    I've used Nagios in the past, but am currently using Zabbix. So far I've been very impressed with it.
     
  7. KillerBunny

    KillerBunny Member

    Joined:
    Jul 14, 2003
    Messages:
    773
    Nagios is great. There's a consulting company who support it in South Australia which helps to. I think it was www.comunet.com.au
     
  8. -=N0N@ME420=-

    -=N0N@ME420=- Member

    Joined:
    May 9, 2004
    Messages:
    7,562
    Juuuuust wondering... how good is extraordinary good ? Rough figure? 10c per license?
     
    Last edited: Jul 4, 2009
  9. fref99

    fref99 Member

    Joined:
    Aug 11, 2002
    Messages:
    839
    Location:
    Switzerland
    Hi,

    I'm using munin on my ubuntu backend server, mainly for long term graphics, the error reporting system is weak but can be linked into Nagios if required.

    Regards
    FREF99
     
  10. didi79

    didi79 Member

    Joined:
    Dec 6, 2003
    Messages:
    34
    We just had 2 interns/trainees deploy an OpenNMS setup for monitoring our servers and network. Works great and they say it's a little bit easier to set up than Nagios.
     
  11. kbekus

    kbekus Member

    Joined:
    Aug 28, 2001
    Messages:
    1,102
    Location:
    Pascoe Vale, Melbourne
    for the local government place I worked at, the Victorian Government negotiated and paid for a 'whole of gov' license that allows departments a significant amount of licenses at no cost.
     
  12. Iceman

    Iceman Member

    Joined:
    Jun 27, 2001
    Messages:
    6,647
    Location:
    Brisbane (nth), Australia
    Speaking of Nagios. I keep seeing Opsview offered as a 'simpler' alternative that comes in a VM for those who can't be bothered messing about with nagios. The big push is the fact that it supports nagios plugins.

    Has anyone tried it?
     
  13. scrantic

    scrantic Member

    Joined:
    Apr 8, 2002
    Messages:
    1,700
    Location:
    3350
  14. j3ll0

    j3ll0 Member

    Joined:
    Jul 13, 2005
    Messages:
    4,695
    Nagviz for the PHBs.

    (Requires NDOUtils and all that entails.)
     
  15. OP
    OP
    Mac

    Mac Member

    Joined:
    Aug 1, 2001
    Messages:
    762
    I believe Server 2008 comes in at under $100. Exchange, ISA and SQL are all similarly priced.
     
  16. OP
    OP
    Mac

    Mac Member

    Joined:
    Aug 1, 2001
    Messages:
    762
    That actually looks really good. Why did you choose to use Zabbix rather than Nagios?
     
  17. zepp

    zepp Member

    Joined:
    Jun 28, 2001
    Messages:
    126
    Location:
    Perth
    I think Zenoss is worth a mention. There's also Hyperic and OpenNMS

    Biggest key for me with these systems is tuning the alerts to improve signal to noise. If it's spitting alerts at you all the time it becomes a bit like the boy who cried wolf.
     
  18. elvis

    elvis Old school old fool

    Joined:
    Jun 27, 2001
    Messages:
    36,261
    Location:
    Brisbane
    Have you considered writing your own?

    They're quite simple to do, and can be done through simple shell scripts (no need to write complex code).

    At the last place I worked we wrote extensive custom Nagios checks. One in particular would check the replication of MySQL databases (we had over 45 production sites, each with a master replicating local data back to head office, and a slave that replicated company-wide data from the head office out to site offices, plus separate DR copies for everything). We wrote a number of checks that would do row counts at each location, and figure out if replication was happening in a timely fashion (or at all).

    It literally took a few hours of effort to knock up, and lowered the resources needed to manually monitor it by both helpdesk and sysadmins.

    That same company had 2 sysadmins, 3 helpdesk and 3 developers who maintained the entire IT operations for 45 sites, with 2000 users, 1600 desktops and 350 servers. The only way we could do that was thanks to Nagios. Without it we would have needed 5 times the people.

    Which ironically is what happened when the company got overtaken by another mob who threw out Nagios, and ended up hiring 5 times the people just to cope with monitoring.
     
  19. username_taken

    username_taken Member

    Joined:
    Oct 19, 2004
    Messages:
    1,352
    Location:
    Austin, TX
    the team was already doing Zabbix when I moved over here, so I just rolled with the punches and after a bit of a learning curve ( it's not intuitive until you understand the base principals, then it's super intuitive ).

    Plenty of plugins available and easy to write your own, at the end of the day probably no better or worse than nagios.

    The other tool I've installed and started to use here is Splunk, basically collates syslogs and gives a simple search engine to them. Using Syslog-NG o push syslogs (and other interesting logs apache,perforce,mysql,etc) for all servers to it.
     
  20. Spingo

    Spingo Member

    Joined:
    Jun 28, 2001
    Messages:
    1,052
    Erm, I do roll my own. In fact, my Windows-based disk checking script also polls various I/O counters via WMI and returns them so that we can generate I/O trending graphs as well as space consumption figures as a percentage (to three decimal places).
     

Share This Page

Advertisement: