d32ca24453
This patch adds the Ceilometer API support back to Aodh. Change-Id: I7b3301fd0569e041218196f1d05047ae7a6138e4
249 lines
9.1 KiB
Python
249 lines
9.1 KiB
Python
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License"); you may
|
|
# not use this file except in compliance with the License. You may obtain
|
|
# a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
|
|
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
|
|
# License for the specific language governing permissions and limitations
|
|
# under the License.
|
|
#
|
|
|
|
from oslo_log import log
|
|
import six
|
|
import stevedore
|
|
|
|
from aodh import evaluator
|
|
from aodh.evaluator import threshold
|
|
|
|
LOG = log.getLogger(__name__)
|
|
|
|
STATE_CHANGE = {evaluator.ALARM: 'outside their threshold.',
|
|
evaluator.OK: 'inside their threshold.',
|
|
evaluator.UNKNOWN: 'state evaluated to unknown.'}
|
|
|
|
|
|
class RuleTarget(object):
|
|
|
|
def __init__(self, rule, rule_evaluator, rule_name):
|
|
self.rule = rule
|
|
self.type = rule.get('type')
|
|
self.rule_evaluator = rule_evaluator
|
|
self.rule_name = rule_name
|
|
self.state = None
|
|
self.trending_state = None
|
|
self.statistics = None
|
|
self.evaluated = False
|
|
|
|
def evaluate(self):
|
|
# Evaluate a sub-rule of composite rule
|
|
if not self.evaluated:
|
|
LOG.debug('Evaluating %(type)s rule: %(rule)s',
|
|
{'type': self.type, 'rule': self.rule})
|
|
try:
|
|
self.state, self.trending_state, self.statistics, __, __ = \
|
|
self.rule_evaluator.evaluate_rule(self.rule)
|
|
except threshold.InsufficientDataError as e:
|
|
self.state = evaluator.UNKNOWN
|
|
self.trending_state = None
|
|
self.statistics = e.statistics
|
|
self.evaluated = True
|
|
|
|
|
|
class RuleEvaluationBase(object):
|
|
def __init__(self, rule_target):
|
|
self.rule_target = rule_target
|
|
|
|
def __str__(self):
|
|
return self.rule_target.rule_name
|
|
|
|
|
|
class OkEvaluation(RuleEvaluationBase):
|
|
|
|
def __bool__(self):
|
|
self.rule_target.evaluate()
|
|
return self.rule_target.state == evaluator.OK
|
|
|
|
__nonzero__ = __bool__
|
|
|
|
|
|
class AlarmEvaluation(RuleEvaluationBase):
|
|
|
|
def __bool__(self):
|
|
self.rule_target.evaluate()
|
|
return self.rule_target.state == evaluator.ALARM
|
|
|
|
__nonzero__ = __bool__
|
|
|
|
|
|
class AndOp(object):
|
|
def __init__(self, rule_targets):
|
|
self.rule_targets = rule_targets
|
|
|
|
def __bool__(self):
|
|
return all(self.rule_targets)
|
|
|
|
def __str__(self):
|
|
return '(' + ' and '.join(six.moves.map(str, self.rule_targets)) + ')'
|
|
|
|
__nonzero__ = __bool__
|
|
|
|
|
|
class OrOp(object):
|
|
def __init__(self, rule_targets):
|
|
self.rule_targets = rule_targets
|
|
|
|
def __bool__(self):
|
|
return any(self.rule_targets)
|
|
|
|
def __str__(self):
|
|
return '(' + ' or '.join(six.moves.map(str, self.rule_targets)) + ')'
|
|
|
|
__nonzero__ = __bool__
|
|
|
|
|
|
class CompositeEvaluator(evaluator.Evaluator):
|
|
def __init__(self, conf):
|
|
super(CompositeEvaluator, self).__init__(conf)
|
|
self.conf = conf
|
|
self._threshold_evaluators = None
|
|
self.rule_targets = []
|
|
self.rule_name_prefix = 'rule'
|
|
self.rule_num = 0
|
|
|
|
@property
|
|
def threshold_evaluators(self):
|
|
if not self._threshold_evaluators:
|
|
threshold_types = ('ceilometer',
|
|
'gnocchi_resources_threshold',
|
|
'gnocchi_aggregation_by_metrics_threshold',
|
|
'gnocchi_aggregation_by_resources_threshold')
|
|
self._threshold_evaluators = stevedore.NamedExtensionManager(
|
|
'aodh.evaluator', threshold_types, invoke_on_load=True,
|
|
invoke_args=(self.conf,))
|
|
return self._threshold_evaluators
|
|
|
|
def _parse_composite_rule(self, alarm_rule):
|
|
"""Parse the composite rule.
|
|
|
|
The composite rule is assembled by sub threshold rules with 'and',
|
|
'or', the form can be nested. e.g. the form of composite rule can be
|
|
like this:
|
|
{
|
|
"and": [threshold_rule0, threshold_rule1,
|
|
{'or': [threshold_rule2, threshold_rule3,
|
|
threshold_rule4, threshold_rule5]}]
|
|
}
|
|
"""
|
|
if (isinstance(alarm_rule, dict) and len(alarm_rule) == 1
|
|
and list(alarm_rule)[0] in ('and', 'or')):
|
|
and_or_key = list(alarm_rule)[0]
|
|
if and_or_key == 'and':
|
|
rules = (self._parse_composite_rule(r) for r in
|
|
alarm_rule['and'])
|
|
rules_alarm, rules_ok = zip(*rules)
|
|
return AndOp(rules_alarm), OrOp(rules_ok)
|
|
else:
|
|
rules = (self._parse_composite_rule(r) for r in
|
|
alarm_rule['or'])
|
|
rules_alarm, rules_ok = zip(*rules)
|
|
return OrOp(rules_alarm), AndOp(rules_ok)
|
|
elif alarm_rule['type'] in self.threshold_evaluators:
|
|
rule_evaluator = self.threshold_evaluators[alarm_rule['type']].obj
|
|
self.rule_num += 1
|
|
name = self.rule_name_prefix + str(self.rule_num)
|
|
rule = RuleTarget(alarm_rule, rule_evaluator, name)
|
|
self.rule_targets.append(rule)
|
|
return AlarmEvaluation(rule), OkEvaluation(rule)
|
|
else:
|
|
LOG.error("Invalid rule type: %s" % alarm_rule['type'])
|
|
return False, False
|
|
|
|
def _reason(self, alarm, new_state, rule_target_alarm):
|
|
transition = alarm.state != new_state
|
|
reason_data = {
|
|
'type': 'composite',
|
|
'composition_form': str(rule_target_alarm)}
|
|
root_cause_rules = {}
|
|
for rule in self.rule_targets:
|
|
if rule.state == new_state:
|
|
root_cause_rules.update({rule.rule_name: rule.rule})
|
|
reason_data.update(causative_rules=root_cause_rules)
|
|
params = {'state': new_state,
|
|
'expression': str(rule_target_alarm),
|
|
'rules': ', '.join(sorted(root_cause_rules)),
|
|
'description': STATE_CHANGE[new_state]}
|
|
if transition:
|
|
reason = (('Composite rule alarm with composition form: '
|
|
'%(expression)s transition to %(state)s, due to '
|
|
'rules: %(rules)s %(description)s') % params)
|
|
|
|
else:
|
|
reason = (('Composite rule alarm with composition form: '
|
|
'%(expression)s remaining as %(state)s, due to '
|
|
'rules: %(rules)s %(description)s') % params)
|
|
|
|
return reason, reason_data
|
|
|
|
def _evaluate_sufficient(self, alarm, rule_target_alarm, rule_target_ok):
|
|
# Some of evaluated rules are unknown states or trending states.
|
|
for rule in self.rule_targets:
|
|
if rule.trending_state is not None:
|
|
if alarm.state == evaluator.UNKNOWN:
|
|
rule.state = rule.trending_state
|
|
elif rule.trending_state == evaluator.ALARM:
|
|
rule.state = evaluator.OK
|
|
elif rule.trending_state == evaluator.OK:
|
|
rule.state = evaluator.ALARM
|
|
else:
|
|
rule.state = alarm.state
|
|
|
|
alarm_triggered = bool(rule_target_alarm)
|
|
if alarm_triggered:
|
|
reason, reason_data = self._reason(alarm, evaluator.ALARM,
|
|
rule_target_alarm)
|
|
self._refresh(alarm, evaluator.ALARM, reason, reason_data)
|
|
return True
|
|
|
|
ok_result = bool(rule_target_ok)
|
|
if ok_result:
|
|
reason, reason_data = self._reason(alarm, evaluator.OK,
|
|
rule_target_alarm)
|
|
self._refresh(alarm, evaluator.OK, reason, reason_data)
|
|
return True
|
|
return False
|
|
|
|
def evaluate(self, alarm):
|
|
if not self.within_time_constraint(alarm):
|
|
LOG.debug('Attempted to evaluate alarm %s, but it is not '
|
|
'within its time constraint.', alarm.alarm_id)
|
|
return
|
|
|
|
LOG.debug("Evaluating composite rule alarm %s ...", alarm.alarm_id)
|
|
self.rule_targets = []
|
|
self.rule_num = 0
|
|
rule_target_alarm, rule_target_ok = self._parse_composite_rule(
|
|
alarm.rule)
|
|
|
|
sufficient = self._evaluate_sufficient(alarm, rule_target_alarm,
|
|
rule_target_ok)
|
|
if not sufficient:
|
|
for rule in self.rule_targets:
|
|
rule.evaluate()
|
|
sufficient = self._evaluate_sufficient(alarm, rule_target_alarm,
|
|
rule_target_ok)
|
|
|
|
if not sufficient:
|
|
# The following unknown situations is like these:
|
|
# 1. 'unknown' and 'alarm'
|
|
# 2. 'unknown' or 'ok'
|
|
reason, reason_data = self._reason(alarm, evaluator.UNKNOWN,
|
|
rule_target_alarm)
|
|
if alarm.state != evaluator.UNKNOWN:
|
|
self._refresh(alarm, evaluator.UNKNOWN, reason, reason_data)
|
|
else:
|
|
LOG.debug(reason)
|