On 25 February 2010 17:55, Ken Foskey <kfos...@tpg.com.au> wrote: > > We all know we should do it. Provide a monitoring system to see how our > system loads are going. I have a couple of links that look interesting: > > http://flapjack-project.com/ > It is local so goes first :-) > Flapjack is a scalable and distributed monitoring system. It natively > talks the Nagios plugin format.
Heh, thanks for the mention. :-) I wouldn't recommend using Flapjack right now unless you want to be testing bleeding edge stuff that is guaranteed to break, or you're a Ruby hacker with an inkling for sysadmin. I'd argue that you're conflating two types of software: statistic collectors (with graphs), and alerters/notifiers. For statistic collection, you cannot go past collectd[0]. collectd is very lightweight (it's written in C), has a plugin architecture (and a boatload of plugins to boot), and is network aware (you can collect stats from all your servers and aggregate them in one place). collectd has a few options for graphing: collection.cgi, collection3.cgi, and Visage[1]. collection*.cgi are CGI scripts (duh) that use RRDtool to generate graphs. Visage draws stats in the browser using JavaScript + SVG. collectd also has a Nagios bridge, so you can plug it into pretty much any alerting/notification system out there. Hope that helps! Lindsay [0] http://collectd.org [1] http://auxesis.github.com/visage (disclaimer: I wrote it) -- w: http://holmwood.id.au/~lindsay/ t: @auxesis -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
