I'm amazed no one has mentioned Sensu on this thread. It certainly deserves a mention. It's a young project, but having used Nagios for 5 years, it's a breath of fresh air, especially with a dynamic cloud based environment. How it works: * Completely ruby based: (gem install sensu). There's a client and server. * You have a master server which publishes checks that clients should run to a queue (rabbitmq). Clients check in to the queue which sets of checks (subscriptions) they should be executing. They also publish results into the queue which the master picks up. * Clients that publish results into the queue are automatically added, no need to add them to the central config. Client deletes can happen via a DELETE call to the API. * Client acknowledgments or check acknowledgments happen via "stashes". * Lots of plugins that work directly with Graphite.
It's still a really young project, but it's much easier to manage than Nagios. It also doesn't quite have the features Nagios has. Not much in the way of automated reporting. References: Introduction to Sensu<https://github.com/joemiller/joemiller.me-intro-to-sensu/blob/master/intro-to-sensu.md> Sensu Checks <https://github.com/sensu/sensu/wiki/Checks> Sensu How to add a check<https://github.com/sensu/sensu/wiki/HOWTO:-Add-a-check> Sensu Handlers <https://github.com/sensu/sensu/wiki/Handlers> Sensu How to add a handler<https://github.com/sensu/sensu/wiki/HOWTO:-Add-a-handler> Sensu Community Plugins <https://github.com/sensu/sensu-community-plugins> Sensu as a collectd replacement<http://petey5king.github.com/2012/03/30/sensu-a-collectd-replacement.html> -kz Kristopher Zentner [email protected] | p: 206-457-2770 On Fri, Mar 22, 2013 at 7:52 AM, Edward Ned Harvey (lopser) < [email protected]> wrote: > I've used zenoss before. Didn't like it. We had problems with the > accuracy of metrics (I think it buffer overflowed or something, getting > disk usage on a several TB volume, reported things like -50% full) ... even > though "technically" it could allow you to create custom metrics via ssh > and so forth, it was confusing, never got that working, etc.**** > > ** ** > > I haven't used any of these others.**** > > ** ** > > Looking at the nagios site, it looks like, you're supposed to install it > on the server you monitor. Installing httpd, mysql, configuring selinux, > etc. Which is not what I want. **** > > ** ** > > I want to install a centralized monitoring / alerting system, and deploy a > tiny little plugin (or something) to each of the systems to be monitored. > The production systems already run apache, mysql, etc, and I don't want any > dependencies on any installation packages to conflict or cause any > disruption to existing production services. If I need to configure httpd > on the system to be monitored, it's a nonstarter.**** > > ** ** > > I primarily care about linux systems (but other OSes are nice to support > too). Want alerts, both predictive and reactionary (notify me if a system > is down, but also notify me when disk usage is over 90% or the CPU stays > over 95% for 10 minutes, or the system begins thrashing swap, etc, so I can > hopefully avoid system down.) etc.**** > > ** ** > > Thanks for suggestions.**** > > _______________________________________________ > Tech mailing list > [email protected] > https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech > This list provided by the League of Professional System Administrators > http://lopsa.org/ > >
_______________________________________________ Tech mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
