Hi Openstack masters. I have ceilometer collecting a bunch of measurements for me, but on a per-server level. I would like to get a periodic snapshot of a project-wide resource utilization; I can use limits for a most of them, except for the networking and disk usages.

My question is - given that ceilometer already collects all this data on a per-vm basis what is the best, or recommended way to get the totals for all VMs? I see couple of possibilities:

1. Aggregate the VM network.incoming.packets.rate/network.outgoing.packets.rate

   • seems a little inaccurate, and in addition openstack metric aggregates command is poorly documented, so coming up with an example aggregation is hard, and cannot groupby timestamp either, but I am open to suggestions here

2. Have a meter on the subnet, which was added to the project for any traffic going in and out.

   • I like this better since it's seems like a much more straightforward thing to do and in addition there is no need to aggregate anything.
   • downside - I could not figure out how to do it so far. Any help here would be appreciated.

Thank you!