allow to dynamically set the "scaling_adjustment" for OS::Heat::ScalingPolicy per workload.

Registered by Jeffrey Guan on 2017-03-17

Issue in current OS::Heat::ScalingPolicy:
In OS::Heat::ScalingPolicy, there is a property named "scaling_adjustment", which is pre-set in hot template. There is no chance for
a system to dynamically set this property according dynamic workload. That is to say, only "scaling_adjustment" number of servers
will be scaled for current ceilometer alarm. If another server is needed, then it should wait another ceilometer alarm after
cooldown time. However, during the cooldown time, the scaled system still cannot handle the over load.

Suggestions:
In order to meet this circumstance, it might be reasonable to allow a dynamic "scaling_adjustment", which is dynamically
set per the incomming workload. For example, if incomming workload needs only one extra server, then "scaling_adjustment"
is set to 1, if incomming workload needs three extra servers , then "scaling_adjustment" is set to 3.

More think:
If dynamic way to set "scaling_adjustment" is allowed, it will also be useful when auto scaling worked with AI or ML.
And it can improve the ability to response changes in system.

Blueprint information

Status:
Not started
Approver:
None
Priority:
Undefined
Drafter:
Jeffrey Guan
Direction:
Needs approval
Assignee:
jeguan
Definition:
New
Series goal:
None
Implementation:
Unknown
Milestone target:
None

Related branches

Sprints

Whiteboard

(?)

Work Items

This blueprint contains Public information 
Everyone can see this information.

Subscribers

No subscribers.