Coverage for watcher/decision_engine/planner/weight.py: 90%
85 statements
« prev ^ index » next coverage.py v7.8.2, created at 2025-06-17 12:22 +0000
1# -*- encoding: utf-8 -*-
2#
3# Authors: Vincent Francoise <Vincent.FRANCOISE@b-com.com>
4# Alexander Chadin <a.chadin@servionica.ru>
5# Licensed under the Apache License, Version 2.0 (the "License");
6# you may not use this file except in compliance with the License.
7# You may obtain a copy of the License at
8#
9# http://www.apache.org/licenses/LICENSE-2.0
10#
11# Unless required by applicable law or agreed to in writing, software
12# distributed under the License is distributed on an "AS IS" BASIS,
13# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
14# implied.
15# See the License for the specific language governing permissions and
16# limitations under the License.
18import collections
20import networkx as nx
21from oslo_config import cfg
22from oslo_log import log
24from watcher.common import utils
25from watcher.decision_engine.planner import base
26from watcher import objects
28LOG = log.getLogger(__name__)
31class WeightPlanner(base.BasePlanner):
32 """Weight planner implementation
34 This implementation builds actions with parents in accordance with weights.
35 Set of actions having a higher weight will be scheduled before
36 the other ones. There are two config options to configure:
37 action_weights and parallelization.
39 *Limitations*
41 - This planner requires to have action_weights and parallelization configs
42 tuned well.
43 """
45 def __init__(self, config):
46 super(WeightPlanner, self).__init__(config)
48 action_weights = {
49 'nop': 70,
50 'volume_migrate': 60,
51 'change_nova_service_state': 50,
52 'sleep': 40,
53 'migrate': 30,
54 'resize': 20,
55 'turn_host_to_acpi_s3_state': 10,
56 'change_node_power_state': 9,
57 }
59 parallelization = {
60 'turn_host_to_acpi_s3_state': 2,
61 'resize': 2,
62 'migrate': 2,
63 'sleep': 1,
64 'change_nova_service_state': 1,
65 'nop': 1,
66 'change_node_power_state': 2,
67 'volume_migrate': 2
68 }
70 @classmethod
71 def get_config_opts(cls):
72 return [
73 cfg.DictOpt(
74 'weights',
75 help="These weights are used to schedule the actions. "
76 "Action Plan will be build in accordance with sets of "
77 "actions ordered by descending weights."
78 "Two action types cannot have the same weight. ",
79 default=cls.action_weights),
80 cfg.DictOpt(
81 'parallelization',
82 help="Number of actions to be run in parallel on a per "
83 "action type basis.",
84 default=cls.parallelization),
85 ]
87 @staticmethod
88 def chunkify(lst, n):
89 """Yield successive n-sized chunks from lst."""
90 n = int(n)
91 if n < 1: 91 ↛ 93line 91 didn't jump to line 93 because the condition on line 91 was never true
92 # Just to make sure the number is valid
93 n = 1
95 # Split a flat list in a list of chunks of size n.
96 # e.g. chunkify([0, 1, 2, 3, 4], 2) -> [[0, 1], [2, 3], [4]]
97 for i in range(0, len(lst), n):
98 yield lst[i:i + n]
100 def compute_action_graph(self, sorted_weighted_actions):
101 reverse_weights = {v: k for k, v in self.config.weights.items()}
102 # leaf_groups contains a list of list of nodes called groups
103 # each group is a set of nodes from which a future node will
104 # branch off (parent nodes).
106 # START --> migrate-1 --> migrate-3
107 # \ \--> resize-1 --> FINISH
108 # \--> migrate-2 -------------/
109 # In the above case migrate-1 will be the only member of the leaf
110 # group that migrate-3 will use as parent group, whereas
111 # resize-1 will have both migrate-2 and migrate-3 in its
112 # parent/leaf group
113 leaf_groups = []
114 action_graph = nx.DiGraph()
115 # We iterate through each action type category (sorted by weight) to
116 # insert them in a Directed Acyclic Graph
117 for idx, (weight, actions) in enumerate(sorted_weighted_actions):
118 action_chunks = self.chunkify(
119 actions, self.config.parallelization[reverse_weights[weight]])
121 # We split the actions into chunks/layers that will have to be
122 # spread across all the available branches of the graph
123 for chunk_idx, actions_chunk in enumerate(action_chunks):
124 for action in actions_chunk:
125 action_graph.add_node(action)
127 # all other actions
128 parent_nodes = []
129 if not idx and not chunk_idx:
130 parent_nodes = []
131 elif leaf_groups: 131 ↛ 134line 131 didn't jump to line 134 because the condition on line 131 was always true
132 parent_nodes = leaf_groups
134 for parent_node in parent_nodes:
135 action_graph.add_edge(parent_node, action)
136 action.parents.append(parent_node.uuid)
138 if leaf_groups:
139 leaf_groups = []
140 leaf_groups.extend([a for a in actions_chunk])
142 return action_graph
144 def schedule(self, context, audit_id, solution):
145 LOG.debug('Creating an action plan for the audit uuid: %s', audit_id)
146 action_plan = self.create_action_plan(context, audit_id, solution)
148 sorted_weighted_actions = self.get_sorted_actions_by_weight(
149 context, action_plan, solution)
150 action_graph = self.compute_action_graph(sorted_weighted_actions)
152 self._create_efficacy_indicators(
153 context, action_plan.id, solution.efficacy_indicators)
155 if len(action_graph.nodes()) == 0:
156 LOG.warning("The action plan is empty")
157 action_plan.state = objects.action_plan.State.SUCCEEDED
158 action_plan.save()
160 self.create_scheduled_actions(action_graph)
161 return action_plan
163 def get_sorted_actions_by_weight(self, context, action_plan, solution):
164 # We need to make them immutable to add them to the graph
165 action_objects = list([
166 objects.Action(
167 context, uuid=utils.generate_uuid(), parents=[],
168 action_plan_id=action_plan.id, **a)
169 for a in solution.actions])
170 # This is a dict of list with each being a weight and the list being
171 # all the actions associated to this weight
172 weighted_actions = collections.defaultdict(list)
173 for action in action_objects:
174 action_weight = self.config.weights[action.action_type]
175 weighted_actions[action_weight].append(action)
177 return reversed(sorted(weighted_actions.items(), key=lambda x: x[0]))
179 def create_scheduled_actions(self, graph):
180 for action in graph.nodes():
181 LOG.debug("Creating the %s in the Watcher database",
182 action.action_type)
183 try:
184 action.create()
185 except Exception as exc:
186 LOG.exception(exc)
187 raise
189 def create_action_plan(self, context, audit_id, solution):
190 strategy = objects.Strategy.get_by_name(
191 context, solution.strategy.name)
193 action_plan_dict = {
194 'uuid': utils.generate_uuid(),
195 'audit_id': audit_id,
196 'strategy_id': strategy.id,
197 'state': objects.action_plan.State.RECOMMENDED,
198 'global_efficacy': solution.global_efficacy,
199 }
201 new_action_plan = objects.ActionPlan(context, **action_plan_dict)
202 new_action_plan.create()
204 return new_action_plan
206 def _create_efficacy_indicators(self, context, action_plan_id, indicators):
207 efficacy_indicators = []
208 for indicator in indicators: 208 ↛ 209line 208 didn't jump to line 209 because the loop on line 208 never started
209 efficacy_indicator_dict = {
210 'uuid': utils.generate_uuid(),
211 'name': indicator.name,
212 'description': indicator.description,
213 'unit': indicator.unit,
214 'value': indicator.value,
215 'action_plan_id': action_plan_id,
216 }
217 new_efficacy_indicator = objects.EfficacyIndicator(
218 context, **efficacy_indicator_dict)
219 new_efficacy_indicator.create()
221 efficacy_indicators.append(new_efficacy_indicator)
222 return efficacy_indicators