On (1), look through the libganglia sources (part of the core installation). Most statistics are gathered as you would expect, but some are harder to get to than just reading /proc or kstat. Also, the Python and SFlow modules add all kinds of other collectors.
For (2), yes, that can be done. If you're very careful and detail oriented (about things like file names, RRD resolutions, etc), you can even keep your old RRD files and display them in both systems, so you have a longer-term historic basis for comparisons. Also: the current state of the grid can always be retrieved from gmetad via XML, which is often handy independent of RRD or the ganglia front-end. What's a bit difficult about (2) is that the information about which metrics have "expired" (actually, the list of currently actively up-to-date metrics) is maintained only by gmetad. In a large grid, you'll always have some machines that are down or some metric you stopped monitoring, so you have to be careful not to present stale information; if you only present the graphs (especially without anchoring the most-recent time), you force your users to read the legends to figure out when the last sample was. Or just check the mtime of the rrd files; if they're older than you think they should be, don't present them, or present them in a different way. -- ReC On Tue, Mar 1, 2011 at 9:26 AM, William Saxton <[email protected]> wrote: > Hi all (potential) new ganglia user here, with a couple quick questions > that I couldn't find the answers to via google. > > 1) Where can I find how ganglia gathers information from a system? I'm > specifically looking to compare how, for example, Solaris gathers its > information (probably through kstat) and Linux gathers its (probably via > /proc) but I'm looking, specifically, for the code that shows this. Is > it available? > > 2) Does anyone have any experience with using ganglia, just as a backend > for storage of RRD data, but then using their own custom front-end? We > are currently using our own home-grown RRD back-end/front-end solution > and, although ganglia looks like a more robust method of data > collection/storage...we're thinking we might just want to port that for > a back-end solution and continue using our own front-end. > > Thanks for your time! > > -Bill > > ------------------------------------------------------------------------------ > Free Software Download: Index, Search & Analyze Logs and other IT data in > Real-Time with Splunk. Collect, index and harness all the fast moving IT data > generated by your applications, servers and devices whether physical, virtual > or in the cloud. Deliver compliance at lower cost and gain new business > insights. http://p.sf.net/sfu/splunk-dev2dev > _______________________________________________ > Ganglia-general mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/ganglia-general > ------------------------------------------------------------------------------ Free Software Download: Index, Search & Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
