I Hate Nagios!

OK gang. There HAS to be something better than Nagios out there. My kingdom for a decent OSS monitoring package.

  1. Try zabbix (http://www.zabbix.org/), it’s at least half decent once you get it running…

  2. What exactly are you trying to monitor?

  3. Have ya tried monit? http://www.tildeslash.com/monit/. It has been working for us so far.

  4. Have ya tried monit? http://www.tildeslash.com/monit/

    It has worked well for us.

  5. Kevin,

    It has no pretty GUI but the tool I’ve always liked has been Mon. It’s old school and not trendy but it does the job and the configs can be managed with (gasp) M4.


  6. It’s tough to beat mon for simplicity and clarity. It seems like some of the big frameworks need to be monitored themselves just to make sure they’re working…

    We also use collectd (http://collectd.org/) for high-resolution stats collection, output as RRD graphs.

    Put sparkline-sized collectd graphs and lightly parsed mon data on a single web page, and you can watch a lot of servers without a lot of complexity.

  7. How is nagios failing you?

  8. What kind of nagios problems are you having? Monit is nice, and I have not tried zabbix. However, Nagios strong sense of its role in monitoring is (in my opinion) highly desirable. It has a clearly defined model and does not confuse: services, the status of services, the dependencies of services, the process of checking services, reporting, and escalation of reporting.

    I think the most typical problem of deploying nagios is that one has to understand the model before configuring the tool. Other monitoring systems that confuse all the things listed above can be deployed quickly but are difficult to extend in the long-term.

  9. Try GroundWork Monitor Open Source, which unifies a bunch of open source monitoring tools, including Nagios, into a single extensible package. It makes configuring and deploying monitoring _much_ easier than straight Nagios.

    New rpms and a bootable ISO have just been uploaded to SourceForge.

  10. Henk Admin

    I don’t really care for GroundWork. For the following reasons:
    You have a distro, like deb or whatever, then you bring this thing in and it brings along it’s own apache, it’s huge bunch of libs, etc etc etc, What the heck is that? (That’s marketing)
    Don’t get me wrong, it works well and looks great, but it’s all about selling something. I like apt and yum and I don’t want another web server and another set of libs that aren’t going to be updated with my package manager. Groundwork takes nagios in the right direction, but then it pushes the user to the cash register. which again, is fine if that’s ok with you.

  11. FWIW, Hyperic also provides a plugin to Nagios offering a seamless and easy upgrade from your Nagios implementation. A nice effect to get you going fast without losing all your hard work implementing a complicated Nagios installation. http://www.hyperic.com.

  12. jack vengance

    We were running Groundwork on a centOS machine, then a more junior level admin attempted to bring in mysql with yum. This sounds like what Henk Admin is talking about. We had an outage until we realized that the conflicting version of mysql had broken Groundwork. I guess Groundwork should sell their own version of Linux like “Evil” Oracle.

  13. Sam Weiss

    Zenoss is another company that is pushing the misleading “Open Source” marketing solution. If an rpm is availble, and the promoting company is in business to sell support the software should be usable after an install.
    Zenoss has numerous problems, including the humorous “can’t login to the interface” issue upon a trial install. There are so many ‘buy it now’ buttons associated with NMS systems that purport to be “Open Source”. No wonder HP and Unisys can still charge huge amounts, there is no easy to setup and install and maintain NMS anywhere. (Pay us and we’ll get it working for you. Pfffff)

%d bloggers like this: