From: Bernard Li <be...@va...> - 2009-09-22 05:27:34
|
P.S. Spike replied to your original email, in case you missed it: http://www.mail-archive.com/gan...@li.../msg05178.html Thanks, Bernard On Mon, Sep 21, 2009 at 10:26 PM, Bernard Li <be...@va...> wrote: > Hi Vladimir: > > Since this is more of a development discussion, I suggest we move this > thread to ganglia-developers. If you haven't subscribed yet, I > strongly encourage you to. > > Thanks for your contributions. > > Cheers, > > Bernard > > On Mon, Sep 21, 2009 at 12:57 PM, Vladimir Vuksan <vl...@ve...> wrote: >> While waiting for a never ending compile to end I decided to hack Ganglia >> myself :-) to do what I proposed. I changed the process_tcp_accept_channel >> to listen for "commands" in the format of >> >> HOSTNAME METRIC >> >> As a result you get something like this >> >> $ echo "localhost.localdomain cpu_nice" | nc localhost 8700 >> <HOST NAME="localhost.localdomain"> >> <METRIC NAME="cpu_nice" VAL="0.0" TYPE="float" UNITS="%" TN="0" TMAX="90" >> DMAX="0" SLOPE="both"> >> <EXTRA_DATA> >> <EXTRA_ELEMENT NAME="GROUP" VAL="cpu"/> >> <EXTRA_ELEMENT NAME="DESC" VAL="Percentage of CPU utilization that occurred >> while executing at the user level with nice priority"/> >> <EXTRA_ELEMENT NAME="TITLE" VAL="cpu_nice"/> >> </EXTRA_DATA> >> </METRIC> >> </HOST> >> $ >> >> Code is not particularly beautiful and likely needs security checks etc. but >> does the job. Obviously this breaks tcp_accept_channel behaviour. Possibly >> it should be run on a separate port or if you don't care about breaking >> legacy behavior it could be run on the same port ie. by having gmetad send >> SEND_ALL command. >> >> Patch is attached. >> >> Vladimir >> >> >> >> On Wed, 16 Sep 2009, Bernard Li wrote: >> >>> +1 on the general idea, not sure about the implementation though. >>> >>> Let's get Spike into this discussion -- yes Spike, I'm calling you out! >>> ;-) >>> >>> Cheers, >>> >>> Bernard >>> >>> On Wed, Sep 16, 2009 at 11:49 AM, Vladimir Vuksan <vl...@ve...> wrote: >>>> >>>> There have been some tweets that someone was working on a REST interface >>>> for Ganglia. At first I thought it wasn't such a big deal but I think >>>> that >>>> adding a simplistic interface to Ganglia would be a nice addition ie. >>>> something like >>>> >>>>> telnet ganglia 8653 >>>> >>>> METRIC web1 load_one >>>> >>>> Which would echo out the current value for load_one. That way you can >>>> avoid parsing out the XML to get those values. I think for large sites it >>>> makes a lot of sense. Granted there are "workarounds" that could be >>>> implemented and people have. >>>> >>>> >>>> I believe all that needs to be done is add a function similar to >>>> >>>> process_tcp_accept_channel >>>> >>>> which doesn't dump out the whole XML but selective value. I would put it >>>> on a separate port call it tcp_get_metric. If you were so inclined you >>>> could even make it a HTTP interface. >>>> >>>> Thoughts ? >>>> >>>> Vladimir >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Come build with us! The BlackBerry® Developer Conference in SF, CA >>>> is the only developer event you need to attend this year. Jumpstart your >>>> developing skills, take BlackBerry mobile applications to market and stay >>>> ahead of the curve. Join us from November 9-12, 2009. Register >>>> now! >>>> http://p.sf.net/sfu/devconf >>>> _______________________________________________ >>>> Ganglia-general mailing list >>>> Gan...@li... >>>> https://lists.sourceforge.net/lists/listinfo/ganglia-general >>>> >> > |