RRDs are updated every time the metadaemon polls an XML source (usually
a monitoring core). This tends to be more frequently than the most
frequently-updated metric.
I'm pretty sure it then walks the entire metric tree and updates every
RRD. It's been a while since I've looked in that code...
Benoit Van Dare wrote:
I have a question about how ganglia reports metrics. I have metric that
is reported every minute for every node in the cluster. When I dump the
contents of the rrd file for that metric for a single node it shows the
value being changed multiple times in a minute. Why is that?
Thanks
Benoit
------------------------------------------------------------------------
Do you Yahoo!?
Free Pop-Up Blocker - Get it now
<http://us.rd.yahoo.com/slv/mailtag/*http://companion.yahoo.com/>
