kingbird/kingbird/engine/README.rst

Service

Kingbird Service has responsibility for:

Monitoring the job/smaller jobs status, and return the result to Kingbird API if needed.

Generate task to purge time-out jobs from Kingbird Database

Multiple Kingbird API could run in parallel, and also can work in multi-worker mode.

Multiple Kingbird Engine will be designed and run in stateless mode, persistent data will be accessed (read and write) from the Kingbird Database through the DAL module.

service.py:

run KB service in multi-worker mode, and establish RPC server

engine_cfg.py:

configuration and initialization for Engine service

quota_manager.py

Manages all the quota related activies such as Periodic Quota Sync, One Demand Quota Sync, Get Total Usage for a Project, Read Kingbird global limit from DB/Conf file etc..

Quota sync happens based on below formula: Global_remaining_limit = Kingbird_global_limit(from DB/Conf) - Su(sum of all usages from all regions) Region_new_limit = Global_remaining_limit + resource_usage_in_that_region.

Reference link: https://etherpad.opnfv.org/p/centralized_quota_management

On Demand Quota Sync: Creates threads for each region and syncs the limits for each quota concurrently.

Periodic Quota Sync: Creates threads for each Project and calls quota sync for project(On Demand Quota sync) for syncing project.

Caches OpenStack region specific clients so reduced traffic.