James
I am not (by far) the most experienced with rrdtool, but I was where you are now back in April (been using mrtg for a few years though) I am currently using rrdtool across hundreds of metrics. The predominant usage is as a graphical web displayer for netiq appmanager collected windows performance data. Very quick web pages showing many servers on the same graph. RRDtool graphing is running across 4 netiq main servers. The web display requests are sent out to all servers. the graphics files then come back to the user. And really fast too... I am also using it for dns response time, web cache traffic, remote access trends, etc... the usual.... A few comments from a relatively recent "newbie" : - read the tutorials from Alex van den Bogaerdt . and then re-read them (heh) - I found that it was best for me to write my own frontend to rrdtool. There was just too many reason that we needed it to look just a certain way. Besides, doing your own graphs causes you to learn the other rrdtool verbs also. - Each server/metric combination is its own rrd file - 5 min for cpu, 60 min for disk. whatever makes sense. - I use png files for my graphics (not gif). my netscape users do not tend to like this though - alarms - this is likely a function of the data collector, not rrdtool, though you can use mrtg fetch to see recent values (sort of expensive though). rrdtool's primary strengths seem to me to be visual. btw - mrtg has alert capability also. - I am using the "rrdtool fetch" to build a way for my automation to "learn" what is normal and then alert when it is not. so... dynamic thresholds - a way to use the rrd files not only for graphics, but also for backend "knowledge" hope this helps Don Mahler SAIC/Telcordia -- Unsubscribe mailto:[EMAIL PROTECTED] Help mailto:[EMAIL PROTECTED] Archive http://www.ee.ethz.ch/~slist/rrd-users WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
