Difference between revisions of "EfficientMetering/GrizzlySummit/StateOfMetering"
| Line 10: | Line 10: | ||
** meter interpretations | ** meter interpretations | ||
** storage | ** storage | ||
| − | |||
** reusable messaging | ** reusable messaging | ||
| + | ** swap out unused parts (turning off meters, using a different system for storing, etc.) | ||
** etc. | ** etc. | ||
* Architecture | * Architecture | ||
Revision as of 18:25, 30 August 2012
State of Metering
Describe the state and direction of the project.
- Project Goals
- metering, not monitoring
- pluggable behaviors
- meter sources
- meter interpretations
- storage
- reusable messaging
- swap out unused parts (turning off meters, using a different system for storing, etc.)
- etc.
- Architecture
- based on Nick’s diagram https://docs.google.com/drawings/pub?id=1_cIFir6HS6jSkPw7chrmyu8DGE2ZgXk79Kbj8nw-Hqo&w=733&h=595
- discuss use of plugins
- cover components
- agents
- collector
- listeners
- db engine plugins
- API basics
- examples of supported queries
- find updated resources
- duration & volume calculations
- raw events
- Implemented Meters
- instance hours (including flavor variation)
- CPU time
- disk I/O
- RAM allocation
- VCPU allocation
- root disk size
- ephemeral disk size
- floating IP allocation
- Grizzly Work To Be Done
- complete notification integration with other projects
- emit more notifications
- emit meter messages directly instead of converting notifications
- more meters
- volume
- network bandwidth
- replace RPC-based meter messages with simpler notifications
- easier for non-OpenStack code to send meter messages
- integrate with Horizon for users to see their data
- Getting the Code
- install via devstack
- installation instructions online at http://ceilometer.readthedocs.org/
- source in github at https://github.com/stackforge/ceilometer
- Questions
