david++; Monitoring != Trending. Yes, Nagios (Zenoss, Zabbix, Groundworks, Opsview, etc) can be made to cough up performance data during checks. Yes, $job gathers that and retains + plots. But Ive yet to encounter an install where such tools are 1:1 substitute for a genuine trending tools (Orca, Munin, Cacti, etc). My $0.02 ...
NB: Splunk seems like a bananas means to correlate data like response time to services based on client requests unless youre sourcing server and client data + munging them together _or_ including response/process time in server log data (which Im not sure can be the canonical source for UX for frontends, etc). Am I being dense here? Not hip to the Splunk special sauce? :) -nick On Wed, Apr 14, 2010 at 7:51 PM, <[email protected]> wrote: > On Wed, 14 Apr 2010, Matt Lawrence wrote: > >> I need to deploy some sort of performance monitoring tools across a few >> hundred systems. In addition to the usual system performance stats, I >> would like to find a set of tools that will easily allow me to collect and >> display various application level information. These are systems doing >> lots of real time transaction processing, so being able to see the graphs >> of what the applications are doing in sync with the system graphs would be >> a very good thing. >> >> The management structure here is that I am on the sysadmin team which does >> not do anything with the applications. So, being able to integrate the >> data collection will make my life easier and make the application support >> folks happier. > > one thing that you need to think about is what you mean by 'monitoring', > there are two branches, and usually tools that do one well don't do the > other as well. > > the first branch is problem detection. > > This is checking if something is out of spec and alerting on it. > > Nagios is a very popular tool for this branch > > this includes doing things like a df to see if your disk is full past a > threshold, or doing a http request to your application to see if it > responds with the correct page (and/or responds within the allowed > deadline) > > > > the second brand is trending/historical data > > This is recording data points and (usually) generating graphs of the data > for analysis. > > rrdtool (the engine that powers MRTG) is a really good tool for this. > > > David Lang > _______________________________________________ > Tech mailing list > [email protected] > http://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] http://lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
