Red Hat Bugzilla – Bug 708376
RfE: report metrics by collection interval
Last modified: 2011-10-06 15:02:03 EDT
We should have a report that lists all schedules by collection interval
interval metric resource
Default sorting would be on interval ascending
[15:42:39] <kmaslowski> hi! RHQ on one of my development server is consuming 100% of one cpu core all the time. I enabled debug and logs stop on entries like: Measurement collection is falling behind... Missed requested time by [30267ms] http://pastebin.com/sXFbDP83
[15:45:42] <pilhuhn> kmaslowski that (indirectly) means that you are collecting too many metrics per minute
[15:46:41] <kmaslowski> so I have enabled too much in metric templates ....
[15:47:29] <pilhuhn> There can be many reasons like the one you mentioned
[15:47:40] <pilhuhn> either too many of them - or at too short intervals
[15:48:07] <pilhuhn> Is your server running at 100% or the agent?
[15:48:35] <kmaslowski> agent
[15:51:33] <pilhuhn> It could also be that the database on the server is too slow and the agent is trying with full power to "get rid" of the collected metrics - mazz?
[15:52:16] <pilhuhn> But the situation should improve when you collect less metrics - either by the number of enabled ones or by the time interval you collect them
[15:52:32] <kmaslowski> how can I find now which metric cause it? :-)
[15:53:42] <kmaslowski> one more question: how can I backup metrics configuration to restore them in the future if the same problem happens?
[15:54:26] <kmaslowski> now I don't remember exacly what I hanged ;) I should document all changes in future :-)
[15:55:07] <pilhuhn> good question(s)