Adam,
Thanks for the response.
Firstly, don't run it in a VM. Observium scales primarily on I/O
throughput. If you have /very/ fast disks you might get away with something that size in a VM.
The Xen servers have a high-end 3par SAN on the backend with the option of moving to faster fibre channel disks as well. I/O seems to be keeping up right now, however I'll keep an eye on it.
Please, please, please do *not* make any changes that won't be committed back into the SVN. Observium is designed to be updated frequently from SVN and has database and other update scripts to make this work. Code changes will break this mechanism. They'll also mean you never update...
I'm willing to try my hand at adding the new OS per the outline in the wiki and submitting patches, or would you prefer to have me provide the data from the devices?
I have the potential of adding several new OS types (two from Cisco's ONS line - 15454 SONET and 15454 MSTP), with a couple of other metro-e aggregation devices that have useful MIBs outside of the normal items.
In practice it makes relatively little difference. Far more impact can be had from increasing disk I/O performance or moving the RRDs into a large ramdisk (i've even had instances where the ramdisk has been on another host due to that being how the RAM was available!)
That's an interesting thought, was the ramdisk in the other host just mounted via NFS?
I have some thoughts ideas on potential features that could be useful
for
other ISP users as well, however I don't want to flood the list right
off
the bat with an even more rambling email.
One of the major things we're missing is good ideas on how to present information, we're especially interested in ideas from other SPs.
I can provide some feed back in a separate thread.
Thanks, Berant