Provides a mechanism to allow the central agent to be horizontally
scaled out, such that each agent polls a disjoint subset of resources.
This is achieved through the use of `tooz` library for distributed
coordination.
If a service wants to use work-load partitioning, it must first
create a PartitionCoordinator object and call its `heartbeat` method
periodically.
To distribute a set of resources over multiple agents, use the
`extract_my_subset` method of the PartitionCoordinator that filters an
iterable, returning only the resources assigned to us.
The `PartitionCoordinator` uses `tooz` to figure out which agents are
in the same group and figures out which resources belong to the
current agent.
DocImpact
Change-Id: I7adef87b03129f4f8b38109bf547c7403cc6adad
Implements: blueprint central-agent-partitioning
9a2f8618de