How to retrieve condensed performance metrics from a whole cluster of ESXi hosts?

Hi,

I am new to the VMware Perl SDK API, but not quite that uninitiated to Perl.

I have written a a few custom Perl Plug-ins for my Nagios and Munin servers, but not yet for monitoring of and retrieving performance stats from our vSphere.

I managed to get the Perl VMware SDK working eventually, after much ado and many curses while wasting time on several failed attempts via RPM package installation,  when I built the whole Perl stack from the sources and had all required Perl modules installed via CPAN.

Still I got annoying SSL warnings because of unknown VMware certificates, because of a broken LWP module release I suppose, that thankfully disappeared after I had updated the libwww-perl to release 5.837.

I would suggest that some distros' package maintainers built working RPMs (only tried RHEL-compatible package managers) that installed with all prerequisites and dependencies right out of the box to save us Linux sysadmins from dependeny hell and messing around with broken packages for hours and even days.

For starters, I installed and configured this Munin plug-in, which builds on the VMware Perl SDK, on my munin master and node to poll our ESXi hosts:

http://foonet.be/2013/06/03/munin-plugin-esxi-cpumemory-usage/

This produced nice Munin charts of memory and cpu usage of each single ESXi host out of 12 that form one cluster.


Although I could combine the fetched data out of 24 RRDs if I tweaked my munin master a little, I suppose, I would like to know if there was a possibly more straight forward approach by e.g. polling our vCenter server and fetch the totals of all 12 cluster nodes by some arcane API method invocation.


Would someone of you savvy SDK hackers know such a handle?


Regards,


Ralph