| To: | Mehran Saliminia <msaliminia@xxxxxxxxx> |
|---|---|
| Subject: | Re: pcp2graphite metrics into percentages |
| From: | fche@xxxxxxxxxx (Frank Ch. Eigler) |
| Date: | Mon, 27 Jul 2015 18:29:55 -0400 |
| Cc: | pcp@xxxxxxxxxxx |
| Delivered-to: | pcp@xxxxxxxxxxx |
| In-reply-to: | <CALsPFxxTA_RHk0K6m0MOBwvT=xXhXpqwbgNBPJM9RAOdfZMJWg@xxxxxxxxxxxxxx> (Mehran Saliminia's message of "Mon, 27 Jul 2015 09:53:10 +0200") |
| References: | <CALsPFxxTA_RHk0K6m0MOBwvT=xXhXpqwbgNBPJM9RAOdfZMJWg@xxxxxxxxxxxxxx> |
| User-agent: | Gnus/5.1008 (Gnus v5.10.8) Emacs/21.4 (gnu/linux) |
msaliminia wrote: > I have been using pcp2graphite. But it sends very large numbers for > the metrics like kernel.all.cpu.idle. Is there any way to convert > this values into percentages before sending them into graphite > server? That metric has "counter" semantics, so normally needs rate-conversion to be more comfortably used. pcp2graphite is not currently capable of rate-conversion, but it turns out that the graphite/whisper system is. So instead of using "FOO.kernel.all.cpu.idle" in your graphite query targets, you can use operators like "perSecond(FOO.kernel.all.cpu.idle)" instead. http://graphite.readthedocs.org/en/latest/functions.html - FChE |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [issue] pmwebd graphite api performance issue, Frank Ch. Eigler |
|---|---|
| Next by Date: | qa/662 failing with python2.7, Ken McDonell |
| Previous by Thread: | pcp2graphite metrics into percentages, Mehran Saliminia |
| Next by Thread: | pcp updates: minor doco and qa, Ken McDonell |
| Indexes: | [Date] [Thread] [Top] [All Lists] |