I'm coming from munin and a CPU graph contains data for system, user, nice, etc ALL on one graph. I just installed ganglia and setup the basic monitoring. It appears that each type of cpu data is a separate graph! WTF is this and can I change the defaults to combine these into a single per host?
That is my question, how do I combine cpu data into a single graph. Also, can I change the layout to something closer to munin's day-week side-by-side layout?
I'm trying to be impartial and give ganglia a chance. ;)
Ganglia is designed for monitoring clusters of many, many machines. It's rather over specified for only a few servers that aren't clustered. I don't know if you can change the default graphs, but you may be better off sticking with a solution designed for regular servers rather than tailored to high performance clusters.
However, if you want to customise Ganglia's graphs, I think you'll need 3.1 and the following web page,
http://sourceforge.net/apps/trac/ganglia/wiki/Custom_graphs
From that page,
Take a look at the summary graphs on the right side: