- 1. Re: [pcp] pcp Digest, Vol 16, Issue 3 (score: 1)
- Author: Jason Rappleye <jason.rappleye@xxxxxxxx>
- Date: Thu, 4 Feb 2010 11:14:45 -0800
- Hi Dave, - Derived metric support. IS ther documentation or a pointer for derived metrics. Been waiting for that. Went digging for it this morning..."man pmRegisterDerived" is a good start, along wit
- /archives/pcp/2010-02/msg00007.html (8,214 bytes)
- 2. Re: [pcp] pcp Digest, Vol 16, Issue 3 (score: 1)
- Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
- Date: Fri, 05 Feb 2010 08:23:52 +1100
- As Jason suggests, man pmRegisterDerived is a good place to start. Here is a small illustration by example, also. $ cat $HOME/.pcp/derived tcp_opens = network.tcp.activeopens + network.tcp.passiveope
- /archives/pcp/2010-02/msg00008.html (9,343 bytes)
- 3. Re: [pcp] pcp Digest, Vol 16, Issue 3 (score: 1)
- Author: David Wright <daw@xxxxxxx>
- Date: Tue, 4 May 2010 13:41:20 -0500 (CDT)
- Ken, Can I use a conditional like if (tcp_opens < 1) then tcp_opens=0
- /archives/pcp/2010-05/msg00004.html (9,757 bytes)
- 4. Re: [pcp] pcp Digest, Vol 16, Issue 3 (score: 1)
- Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
- Date: Wed, 05 May 2010 06:45:53 +1000
- Afraid not Dave. The derived metric expressions are limited to simple arithmetic (albeit over set-valued operands) and a handful of set -> unary value functions - avg(), count(), max(), min(), sum(),
- /archives/pcp/2010-05/msg00005.html (9,228 bytes)
- 5. Re: [pcp] pcp Digest, Vol 16, Issue 3 (score: 1)
- Author: David Wright <daw@xxxxxxx>
- Date: Tue, 4 May 2010 18:02:37 -0500 (CDT)
- I was looking for a way to avoid a negative. pmchart is going funny with my stack chart and flipping to just the derived. mem.util.ncache_clean = mem.util.cache_clean - mem.util.shmem mem.util.mmapp
- /archives/pcp/2010-05/msg00008.html (8,342 bytes)
- 6. Re: [pcp] pcp Digest, Vol 16, Issue 3 (score: 1)
- Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
- Date: Fri, 07 May 2010 07:44:03 +1000
- Does the value of the derived metric go just a little negative (due to some rounding or timing issues in the calculation) or a lot negative (suggesting the semantics of the operand metrics or the der
- /archives/pcp/2010-05/msg00029.html (8,312 bytes)
This search system is powered by
Namazu