Difference between revisions of "Telemetry/RoadMap"
Gordon chung (talk | contribs) (→Road Map) |
Gordon chung (talk | contribs) (→Gnocchi (metric storage)) |
||
| Line 39: | Line 39: | ||
* benchmark rewrite, generating load with python is not efficient enough for good benchmarking | * benchmark rewrite, generating load with python is not efficient enough for good benchmarking | ||
* project unique resource_id | * project unique resource_id | ||
| + | * fair scheduling | ||
| + | * dynamic granularity aggregates | ||
* [https://bugs.launchpad.net/gnocchi/+bugs?field.searchtext=&orderby=-importance&search=Search&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.importance%3Alist=WISHLIST&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch.used=&field.has_branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on wishlist items] | * [https://bugs.launchpad.net/gnocchi/+bugs?field.searchtext=&orderby=-importance&search=Search&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.importance%3Alist=WISHLIST&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch.used=&field.has_branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on wishlist items] | ||
* migrate tempest plugin | * migrate tempest plugin | ||
Revision as of 13:26, 2 November 2016
Contents
Road Map
NOTE: We do not use blueprints in Telemetry projects. If a patch is complex or controversial, please raise the topic to the mailing list using the [telemetry] tag. In very rare cases will a spec or blueprint be required.
Current (Ocata) Cycle Targets: https://blueprints.launchpad.net/ceilometer/ocata
- items from summit - https://etherpad.openstack.org/p/mitaka-telemetry-todos
- pre-existing blueprints - https://blueprints.launchpad.net/ceilometer
Priority Work Items
The following are high priority items targeted for the current cycle that require owners:
Open Work Items
The following is a list of work items that have been approved conceptually and can be targeted for the current development cycle. If a new resource is found, the following will happen; if not, it might take a while.
For more information on a item, contact us on openstack-dev mailing list or on freenode at #openstack-telemetry. The current list of core contributors can be found at each projects respective page:
- https://launchpad.net/~aodh-drivers/+members#active
- https://launchpad.net/~ceilometer-drivers/+members#active
- https://launchpad.net/~gnocchi-drivers/+members#active
Aodh (alarming)
- in-tree functional tests
- event alarm multiple workers
- use oslo.messaging batch notification listener
- adopt oslo.db pagination
Ceilometer (data collection)
- Polling schema - separate polling logic from pipeline
- new polling definition file https://etherpad.openstack.org/p/mitaka-telemetry-polling
- multiple meter/event definition file support
- refine polling
- drop metadata polling requirements
- deprecate ceilometer-api
- generic notification agent
- rally tests
Gnocchi (metric storage)
- indexer sharding support
- benchmark rewrite, generating load with python is not efficient enough for good benchmarking
- project unique resource_id
- fair scheduling
- dynamic granularity aggregates
- wishlist items
- migrate tempest plugin
Future Targets
These ideas have been discussed/introduced at a high-level. They are areas of interests for the users/developers of Ceilometer but still require discussion. Proposals for concrete work items related to these items are welcomed for future cycles:
- log processing
- application level monitoring
- group based configuration control - connect tooz to manage configuration and allow for better control of configuration updates.
- nova cell metrics
- nova availability zone metrics
- project level pipeline control. enable ability to poll at different frequencies per project.
- granular ttl control (project level / meter level)
- CADF event indexer - generic definition to index all CADF events
- stop using WSME in aodh
- Use specific configuration to load selected stevedore extensions, not all of them
- remove stevedore plugins out of tree
- Add more complete evaluation of the upgrade process in the grenade plugin.
- Notification and pollster driven functional/integration tests.
