we are running serveral OpenNebula Instances within our Datacenters.
Our current capacity management is built on top of some Excel Sheets which show us
REAL vs. ALLOCATED vs. USED. This is odd.
All other Services (running on top of the OpenNebula IaaS-Stack) have their own Graphite Metrics collected and pushed them to shiny and cool Dashboards (like Dashing.io).
I think the same solutions (for example ONE Graphite-Metrics out of the Box) for OpenNebula will be quite nice.
It could be a shiny solution for providing long-term statistics and a base for Capacity management.
What do you think? What do you use for capacity management?
For example, a small patch of
im/kvm-probes.d/kvm.rb with some lines of code could generate the Metrics for these kind of statistics? But will this be the right place to gather this stats? Are there any other sources?