In this changeset, I fixed the issue with the basic server
consolidation strategy to now loop over all compute nodes
as expected instead of stopping after the first one.
Change-Id: If594f0df41e39dfb0ef8f0fce41822018490c4ec
Closes-bug: #1548874
This patch set allows to use audit parameters for
workload-stabilization strategy and makes some little
refactoring.
Closes-Bug: #1620604
Change-Id: I60e34611d4dd001beed31666fd11d2ab11c1723c
In this changeset, I implemented a small GMR plugin that converts
the cluster data model structure into an XML structure.
Change-Id: I75548952635a0aa3c7dbd6d068831b5765a5db1a
Closes-Bug: #1620551
In this changeset, I implemented the notification handling (Rx only)
system for consuming incoming notifications, more especially the Nova
ones. The notifications handlers also contain the logic which
incrementally updates the Compute model.
Change-Id: Ia036a5a2be6caa64b7f180de38821b57c624300c
Partially-implements: blueprint cluster-model-objects-wrapper
This change is adding the main logic for the scoring module,
defines entry points for the scoring engine plugins and provides
a watcher-sync tool to enable Watcher database synchronization
without needing to restart any Watcher service.
Partially-Implements: blueprint scoring-module
Change-Id: If10daae969ec27a7008af5173359992e957dcd5e
In this changeset, I implemented the logic which cancels
any audit or action plan whose goal has been re-synced
(upon restarting the Decision Engine).
Partially Implements: blueprint efficacy-indicator
Change-Id: I95d2739eb552d4a7a02c822b11844591008f648e
This patch set renames 'OS-EXT-STS:instance_state' to 'OS-EXT-STS:vm_state'
for correct working of decision_engine scheduler.
Change-Id: I20805a079a991d5f3b8565f52d5f7280c2389bee
Closes-Bug: #1614511
In this changeset, I refactored the whole Watcher codebase to
adopt a naming convention about the various elements of the
Compute model so that it reflects the same naming convention
adopted by Nova.
Change-Id: I28adba5e1f27175f025330417b072686134d5f51
Partially-Implements: blueprint cluster-model-objects-wrapper
Many places, there are more specific asserts which can be used.
I replaced the generic assert with more specific ones, where
it was possible.
This change enhances readibility, and on fail, more useful
message is displayed
Change-Id: I86a6baeae2cd36610a2be10ae5085555246c368a
In this changeset, I merged the metrics_engine package into
the decision_engine one alongside the required changes to make
the tests pass.
Change-Id: Iac1cd266a854212bf4fa8b21c744b076c3b834a8
Partially-Implements: blueprint cluster-model-objects-wrapper
In this changeset, I implemented a background scheduler service
for Watcher and more particularly for the Decision Engine where
I made it create 2 types of job per cluster data model collector
plugin:
- An initial job that is asynchronously executed upon starting the
Decision Engine
- A periodical job that gets triggered every configurable interval
of time
Change-Id: I3f5442f81933a19565217b894bd86c186e339762
Partially-Implements: bluprint cluster-model-objects-wrapper
In this changeset, I made BaseClusterDataModelCollector instances
pluggable. This corresponds to "part 1" of the work items detailed
in the specifications.
Change-Id: Iab1c7e264add9e2cbbbb767e3fd6e99a0c22c691
Partially-Implements: blueprint cluster-model-objects-wrapper
Modifying the api controller for audit objects to allow
creation of audit objects by specifying either an
audit_template uuid/id and/or a goal_id.
strategy_id is optional.
Partially Implements: blueprint persistent-audit-parameters
Change-Id: I7b3eae4d0752a11208f5f92ee13ab1018d8521ad
use parameters to set the threshold for strategies
of Uniform Airflow,Workload balance.
Change-Id: I5a547929eb1e2413468e9a5de16d3fd42cabadf9
Close-bug: #1600707
The hacking checks enforce during the pep8 run functional validations of
the code to ensure deeper filters and code consistency. This change set
adds the hacking checks to the wathcer project. These checks were
seeded from the neutron project, which had a good set of base defaults.
This change set also updates the watcher project to be compliant with
these new hacking checks.
Change-Id: I6f4566d384a7400bddf228aa127a53e6ecc82c2e
This patch set adds implementation for CONTINUOUS type
of audit.
Change-Id: I5f4ec97b2082c8a6b3ccebe36b2a343fa4a67d19
Implements: blueprint continuously-optimization
Strategy provides parameters to customize algorithm behavior. End user
could query then specify parameters for their requirements.
Change-Id: Id097db5f6e79c94b57674c8e5d55b06098abf18c
Implements-bp: optimization-threshold
Currently we have to amend the code of the Planner
for each new added Watcher Action. This patch set will
modifying the DefaultPlanner class and leverage the
plugins-parameters blueprint so that we can configure
weights via the configuration file.
Documentation and unittest also be update in this patch.
Co-Authored-By: Vincent Françoise <vincent.francoise@b-com.com>
Change-Id: Ib794488fcafd1e153a7d7b1f7253686136501872
blueprint: configurable-weights-default-planner
This is one of the algorithm of Intel thermal POC.
It's based on the Airflow of hypervisors.
Change-Id: I48bbab75b20129d85b1fb15ede823474546fe399
blueprint: uniform-airflow-migration-strategy
Refactor all plugin loaders in watcher/applier into
watcher/applier/loading/default.py.
Change-Id: I0f0a87d4f72ead45d34aca1b14219fd2ede42a6f
Closes-Bug: #1591095
This patch removes import of classes and replaces them with import of modules.
Change-Id: Id2502cf96ac7227cf1036cd54a74f3d7acd83479
Closes-Bug: #1594434
The VM workload consolidation strategy sends 'down' instead
of 'disabled' to change nova-compute service state. This patch
will correct it. The same applies for enabling nova-compute service.
Change-Id: I257411ef711b215bd9b56d0bf0479c79a9ef61d8
Closes-Bug: #1591901
This patch fixes the StrategyContext to use the optional
attribute strategy_id.
Change-Id: Ib78581f564282de6cfc7f07495c846615ec1866a
Closed-bug: #1590357
I this changeset, I added the efficacy indicators both at the DB
and at the API level alongside the associated logic.
Partially Implements: blueprint efficacy-indicator
Change-Id: I824553637621da67966103c1b0c01348b09bd836
In this changeset, I broke down the execute() method to sequentially
call 3 methods:
- pre_execute()
- do_execute()
- post_execute()
This changeset also removes the cluster model parameter from the
execute() method to now become a `model` property of a strategy which
is lazy loaded whenever needed.
Partially Implements: blueprint efficacy-indicator
Change-Id: I2f697938db693acfa95b2c2fbecfdc1b733c93fd
In this changeset, I added the new Efficacy, EfficacySpecification
and IndicatorSpecification classes which are the main components for
computing an efficacy.
Partially Implements: blueprint efficacy-indicator
Change-Id: I3a1d62569de2dd6bb6f9a52f6058313fa2b886ce
In this changeset, I decoupled the notion of Goal from the Strategy
by making it a distinct object. Goals are plugins that can be loaded
just like for the strategies.
Partially Implements: blueprint efficacy-indicator
Change-Id: I4378dccd508170b305aa968843228bbc8af78895
This patch set removes normalizing for hardware.cpu.util meter
since the values that comes from ceilometer hardware.cpu.util
are already normalized.
Closes-Bug: #1588257
Change-Id: I9494f2cc9bbaa6dfd168fb515f679eb6d7f2398a
This is one of the algorithm of Intel thermal POC.
It's based on the VM workloads of hypervisors.
Change-Id: I45ab0cf0f05786e6f68025bdd315f38381900a68
blueprint: workload-balance-migration-strategy
