Add doc for workload-stabilization spec
This patch set adds how-to-use documentation for Watcher Overload standard deviation algorithm. Change-Id: I75d7cd0ff8507ca70efb6d9668ae9fbf651a7f97
This commit is contained in:
Alexander Chadin
parent
ad7ae3e676
commit
9761207d8a
@ -91,6 +91,12 @@ Efficacy Indicator
|
||||
|
||||
Add here the Efficacy indicator computed by your strategy.
|
||||
|
||||
Algorithm
|
||||
---------
|
||||
|
||||
Add here either the description of your algorithm or
|
||||
link to the existing description.
|
||||
|
||||
How to use it ?
|
||||
---------------
|
||||
|
||||
|
||||
131
doc/source/strategies/workload-stabilization.rst
Normal file
131
doc/source/strategies/workload-stabilization.rst
Normal file
@ -0,0 +1,131 @@
|
||||
=============================================
|
||||
Watcher Overload standard deviation algorithm
|
||||
=============================================
|
||||
|
||||
Synopsis
|
||||
--------
|
||||
|
||||
**display name**: ``workload_stabilization``
|
||||
|
||||
**goal**: ``workload_balancing``
|
||||
|
||||
.. watcher-term:: watcher.decision_engine.strategy.strategies.workload_stabilization
|
||||
|
||||
Requirements
|
||||
------------
|
||||
|
||||
Metrics
|
||||
*******
|
||||
|
||||
The *workload_stabilization* strategy requires the following metrics:
|
||||
|
||||
============================ ============ ======= =======
|
||||
metric service name plugins comment
|
||||
============================ ============ ======= =======
|
||||
``compute.node.cpu.percent`` ceilometer_ none
|
||||
``hardware.memory.used`` ceilometer_ SNMP_
|
||||
``cpu_util`` ceilometer_ none
|
||||
``memory.resident`` ceilometer_ none
|
||||
============================ ============ ======= =======
|
||||
|
||||
.. _ceilometer: http://docs.openstack.org/admin-guide/telemetry-measurements.html#openstack-compute
|
||||
.. _SNMP: http://docs.openstack.org/admin-guide/telemetry-measurements.html
|
||||
|
||||
Cluster data model
|
||||
******************
|
||||
|
||||
Default Watcher's Compute cluster data model:
|
||||
|
||||
.. watcher-term:: watcher.decision_engine.model.collector.nova.NovaClusterDataModelCollector
|
||||
|
||||
Actions
|
||||
*******
|
||||
|
||||
Default Watcher's actions:
|
||||
|
||||
|
||||
.. list-table::
|
||||
:widths: 30 30
|
||||
:header-rows: 1
|
||||
|
||||
* - action
|
||||
- description
|
||||
* - ``migration``
|
||||
- .. watcher-term:: watcher.applier.actions.migration.Migrate
|
||||
|
||||
Planner
|
||||
*******
|
||||
|
||||
Default Watcher's planner:
|
||||
|
||||
.. watcher-term:: watcher.decision_engine.planner.default.DefaultPlanner
|
||||
|
||||
Configuration
|
||||
-------------
|
||||
|
||||
Strategy parameters are:
|
||||
|
||||
==================== ====== ===================== =============================
|
||||
parameter type default Value description
|
||||
==================== ====== ===================== =============================
|
||||
``metrics`` array |metrics| Metrics used as rates of
|
||||
cluster loads.
|
||||
``thresholds`` object |thresholds| Dict where key is a metric
|
||||
and value is a trigger value.
|
||||
|
||||
``weights`` object |weights| These weights used to
|
||||
calculate common standard
|
||||
deviation. Name of weight
|
||||
contains meter name and
|
||||
_weight suffix.
|
||||
``instance_metrics`` object |instance_metrics| Mapping to get hardware
|
||||
statistics using instance
|
||||
metrics.
|
||||
``host_choice`` string retry Method of host's choice.
|
||||
There are cycle, retry and
|
||||
fullsearch methods. Cycle
|
||||
will iterate hosts in cycle.
|
||||
Retry will get some hosts
|
||||
random (count defined in
|
||||
retry_count option).
|
||||
Fullsearch will return each
|
||||
host from list.
|
||||
``retry_count`` number 1 Count of random returned
|
||||
hosts.
|
||||
==================== ====== ===================== =============================
|
||||
|
||||
.. |metrics| replace:: ["cpu_util", "memory.resident"]
|
||||
.. |thresholds| replace:: {"cpu_util": 0.2, "memory.resident": 0.2}
|
||||
.. |weights| replace:: {"cpu_util_weight": 1.0, "memory.resident_weight": 1.0}
|
||||
.. |instance_metrics| replace:: {"cpu_util": "hardware.cpu.util", "memory.resident": "hardware.memory.used"}
|
||||
|
||||
Efficacy Indicator
|
||||
------------------
|
||||
|
||||
.. watcher-func::
|
||||
:format: literal_block
|
||||
|
||||
watcher.decision_engine.goal.efficacy.specs.ServerConsolidation.get_global_efficacy_indicator
|
||||
|
||||
Algorithm
|
||||
---------
|
||||
|
||||
You can find description of overload algorithm and role of standard deviation
|
||||
here: https://specs.openstack.org/openstack/watcher-specs/specs/newton/implemented/sd-strategy.html
|
||||
|
||||
How to use it ?
|
||||
---------------
|
||||
|
||||
.. code-block:: shell
|
||||
|
||||
$ openstack optimize audittemplate create \
|
||||
at1 workload_balancing --strategy workload_stabilization
|
||||
|
||||
$ openstack optimize audit create -a at1 \
|
||||
-p thresholds='{"memory.resident": 0.05}' \
|
||||
-p metrics='["memory.resident"]'
|
||||
|
||||
External Links
|
||||
--------------
|
||||
|
||||
- `Watcher Overload standard deviation algorithm spec <https://specs.openstack.org/openstack/watcher-specs/specs/newton/implemented/sd-strategy.html>`_
|
||||
@ -16,6 +16,16 @@
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
#
|
||||
"""
|
||||
*Workload Stabilization control using live migration*
|
||||
|
||||
This is workload stabilization strategy based on standard deviation
|
||||
algorithm. The goal is to determine if there is an overload in a cluster
|
||||
and respond to it by migrating VMs to stabilize the cluster.
|
||||
|
||||
It assumes that live migrations are possible in your cluster.
|
||||
|
||||
"""
|
||||
|
||||
import copy
|
||||
import itertools
|
||||
@ -48,35 +58,18 @@ def _set_memoize(conf):
|
||||
|
||||
|
||||
class WorkloadStabilization(base.WorkloadStabilizationBaseStrategy):
|
||||
"""Workload Stabilization control using live migration
|
||||
|
||||
*Description*
|
||||
|
||||
This is workload stabilization strategy based on standard deviation
|
||||
algorithm. The goal is to determine if there is an overload in a cluster
|
||||
and respond to it by migrating VMs to stabilize the cluster.
|
||||
|
||||
*Requirements*
|
||||
|
||||
* Software: Ceilometer component ceilometer-compute running
|
||||
in each compute host, and Ceilometer API can report such telemetries
|
||||
``memory.resident`` and ``cpu_util`` successfully.
|
||||
* You must have at least 2 physical compute nodes to run this strategy.
|
||||
|
||||
*Limitations*
|
||||
|
||||
- It assume that live migrations are possible
|
||||
- Load on the system is sufficiently stable.
|
||||
|
||||
*Spec URL*
|
||||
|
||||
https://review.openstack.org/#/c/286153/
|
||||
"""
|
||||
"""Workload Stabilization control using live migration"""
|
||||
|
||||
MIGRATION = "migrate"
|
||||
MEMOIZE = _set_memoize(CONF)
|
||||
|
||||
def __init__(self, config, osc=None):
|
||||
"""Workload Stabilization control using live migration
|
||||
|
||||
:param config: A mapping containing the configuration of this strategy
|
||||
:type config: :py:class:`~.Struct` instance
|
||||
:param osc: :py:class:`~.OpenStackClients` instance
|
||||
"""
|
||||
super(WorkloadStabilization, self).__init__(config, osc)
|
||||
self._ceilometer = None
|
||||
self._nova = None
|
||||
|
||||
Loading…
Reference in New Issue
Block a user