Peter,
I think you might want to move your questions/discussion to the
developers list.
The thing you have to take in account is, all that gmetric does is
'insert' extra (XML) custom values into the running gmond.
The transmission of this XML is still handled by the gmond, and might or
might not have a delay before the XML is actually seen/placed onto the
XML channel, depending on your setup and machines. I believe gmond does
not transmit data every second by default.
Also you have to think about where you want to parse the XML from.
If you parse it by reading the tcp port of gmetad, you have to remember
that gmetad only updates his XML from data_source's on a interval set in
the config (default 10 seconds).
You might want to dig into the inner workings of Ganglia some more
before you start the project, to see if what you want is possible.
I have had similar issues while developing a (cluster) job
listing/archiving addon for Ganglia, and one of the things I did is
timestamp all the (g)metric's and use a sort of extra
heartbeat/handshake like metric.
Kind regards,
- Ramon.
Peter Bodik wrote:
Does anybody have experience with such setup? Basically,
the number of
hosts would be pretty small, but the number of metrics
could be large
(100s or
1000s) and for some of them we'll need very fast updates.
While it would be nice to have per-process or per-application
metrics, I believe it was recently stated in an e-mail just
sent in the last 24 hours about a similar task that it's not
currently available in Ganglia, although it might be a nice
feature to put into the libmetrics codes.
We're planning to add all the application-specific metrics through gmetric.
So the J2EE app would call gmetric every time it wants to update. For
example, I would create a metric 'EJB_componentX_errorrate' for each
componentX and report the value from J2EE.
So it certainly is possible, but I'm just worried (for example) that some
data might get lost because of such frequent updates or something.
The user-defined metrics are less-efficient, right? So having 1000s (from
each host) could cause some problems ... I'd like to know what kind of
problems we can expect and what do to about them.
Peter
-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Ganglia-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
.--------------------------------------------.
| ing. Ramon Bastiaans |
| HPC - Systems Programmer |
|--------------------------------------------|
| SARA - Computing and Networking Services |
| Kruislaan 415 PO Box 194613 |
| 1098 SJ Amsterdam 1090 GP Amsterdam |
|--------------------------------------------|
| Mail: bastiaans ( a t ) sara ( d o t ) nl |
| Web: http://www.sara.nl/ |
| Phone: +31 (0)20 592 80 19 |
| Fax: +31 (0)20 668 31 67 |
`--------------------------------------------'
