OpenStack Telemetry (Ceilometer)
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

2642 lines
96 KiB

# Copyright 2012 New Dream Network, LLC (DreamHost)
# Copyright 2013 IBM Corp.
# Copyright 2013 eNovance <>
# Copyright Ericsson AB 2013. All rights reserved
# Authors: Doug Hellmann <>
# Angus Salkeld <>
# Eoghan Glynn <>
# Julien Danjou <>
# Ildiko Vancsa <>
# Balazs Gibizer <>
# 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
# 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.
"""Version 2 of the API.
import ast
import base64
import copy
import datetime
import functools
import inspect
import json
import uuid
import croniter
import jsonschema
from oslo.config import cfg
from oslo.utils import netutils
from oslo.utils import strutils
from oslo.utils import timeutils
import pecan
from pecan import rest
import pytz
import six
import wsme
from wsme import types as wtypes
import wsmeext.pecan as wsme_pecan
import ceilometer
from ceilometer.alarm import service as alarm_service
from import models as alarm_models
from ceilometer.api import acl
from import models as event_models
from ceilometer import messaging
from ceilometer.openstack.common import context
from ceilometer.openstack.common.gettextutils import _
from ceilometer.openstack.common import log
from ceilometer import sample
from ceilometer import storage
from ceilometer import utils
LOG = log.getLogger(__name__)
help='Record alarm change events.'
help='Maximum number of alarms defined for a user.'
help='Maximum number of alarms defined for a project.'
cfg.CONF.register_opts(ALARM_API_OPTS, group='alarm')
state_kind = ["ok", "alarm", "insufficient data"]
state_kind_enum = wtypes.Enum(str, *state_kind)
operation_kind = ('lt', 'le', 'eq', 'ne', 'ge', 'gt')
operation_kind_enum = wtypes.Enum(str, *operation_kind)
class ClientSideError(wsme.exc.ClientSideError):
def __init__(self, error, status_code=400):
pecan.response.translatable_error = error
super(ClientSideError, self).__init__(error, status_code)
class EntityNotFound(ClientSideError):
def __init__(self, entity, id):
super(EntityNotFound, self).__init__(
_("%(entity)s %(id)s Not Found") % {'entity': entity,
'id': id},
class AlarmNotFound(ClientSideError):
def __init__(self, alarm, auth_project):
if not auth_project:
msg = _('Alarm %s not found') % alarm
msg = _('Alarm %(alarm_id)s not found in project %'
'(project)s') % {
'alarm_id': alarm, 'project': auth_project}
super(AlarmNotFound, self).__init__(msg, status_code=404)
class OverQuota(ClientSideError):
def __init__(self, data):
d = {
'u': data.user_id,
'p': data.project_id
super(OverQuota, self).__init__(
_("Alarm quota exceeded for user %(u)s on project %(p)s") % d,
def is_over_quota(conn, project_id, user_id):
"""Returns False if an alarm is within the set quotas, True otherwise.
:param conn: a backend connection object
:param project_id: the ID of the project setting the alarm
:param user_id: the ID of the user setting the alarm
over_quota = False
# Start by checking for user quota
user_alarm_quota = cfg.CONF.alarm.user_alarm_quota
if user_alarm_quota is not None:
user_alarms = list(conn.get_alarms(user=user_id))
over_quota = len(user_alarms) >= user_alarm_quota
# If the user quota isn't reached, we check for the project quota
if not over_quota:
project_alarm_quota = cfg.CONF.alarm.project_alarm_quota
if project_alarm_quota is not None:
project_alarms = list(conn.get_alarms(project=project_id))
over_quota = len(project_alarms) >= project_alarm_quota
return over_quota
class AdvEnum(wtypes.wsproperty):
"""Handle default and mandatory for wtypes.Enum."""
def __init__(self, name, *args, **kwargs):
self._name = '_advenum_%s' % name
self._default = kwargs.pop('default', None)
mandatory = kwargs.pop('mandatory', False)
enum = wtypes.Enum(*args, **kwargs)
super(AdvEnum, self).__init__(datatype=enum, fget=self._get,
fset=self._set, mandatory=mandatory)
def _get(self, parent):
if hasattr(parent, self._name):
value = getattr(parent, self._name)
return value or self._default
return self._default
def _set(self, parent, value):
if self.datatype.validate(value):
setattr(parent, self._name, value)
class CronType(wtypes.UserType):
"""A user type that represents a cron format."""
basetype = six.string_types
name = 'cron'
def validate(value):
# raises ValueError if invalid
return value
class _Base(wtypes.Base):
def from_db_model(cls, m):
return cls(**(m.as_dict()))
def from_db_and_links(cls, m, links):
return cls(links=links, **(m.as_dict()))
def as_dict(self, db_model):
valid_keys = inspect.getargspec(db_model.__init__)[0]
if 'self' in valid_keys:
return self.as_dict_from_keys(valid_keys)
def as_dict_from_keys(self, keys):
return dict((k, getattr(self, k))
for k in keys
if hasattr(self, k) and
getattr(self, k) != wsme.Unset)
class Link(_Base):
"""A link representation."""
href = wtypes.text
"The url of a link"
rel = wtypes.text
"The name of a link"
def sample(cls):
return cls(href=('http://localhost:8777/v2/meters/volume?'
class Query(_Base):
"""Query filter."""
# The data types supported by the query.
_supported_types = ['integer', 'float', 'string', 'boolean']
# Functions to convert the data field to the correct type.
_type_converters = {'integer': int,
'float': float,
'boolean': functools.partial(
strutils.bool_from_string, strict=True),
'string': six.text_type,
'datetime': timeutils.parse_isotime}
_op = None # provide a default
def get_op(self):
return self._op or 'eq'
def set_op(self, value):
self._op = value
field = wtypes.text
"The name of the field to test"
# op = wsme.wsattr(operation_kind, default='eq')
# this ^ doesn't seem to work.
op = wsme.wsproperty(operation_kind_enum, get_op, set_op)
"The comparison operator. Defaults to 'eq'."
value = wtypes.text
"The value to compare against the stored data"
type = wtypes.text
"The data type of value to compare against the stored data"
def __repr__(self):
# for logging calls
return '<Query %r %s %r %s>' % (self.field,
def sample(cls):
return cls(field='resource_id',
def as_dict(self):
return self.as_dict_from_keys(['field', 'op', 'type', 'value'])
def _get_value_as_type(self, forced_type=None):
"""Convert metadata value to the specified data type.
This method is called during metadata query to help convert the
querying metadata to the data type specified by user. If there is no
data type given, the metadata will be parsed by ast.literal_eval to
try to do a smart converting.
NOTE (flwang) Using "_" as prefix to avoid an InvocationError raised
from wsmeext/ It's OK to call it outside the Query class.
Because the "public" side of that class is actually the outside of the
API, and the "private" side is the API implementation. The method is
only used in the API implementation, so it's OK.
:returns: metadata value converted with the specified data type.
type = forced_type or self.type
converted_value = self.value
if not type:
converted_value = ast.literal_eval(self.value)
except (ValueError, SyntaxError):
# Unable to convert the metadata value automatically
# let it default to self.value
if type not in self._supported_types:
# Types must be explicitly declared so the
# correct type converter may be used. Subclasses
# of Query may define _supported_types and
# _type_converters to define their own types.
raise TypeError()
converted_value = self._type_converters[type](self.value)
except ValueError:
msg = (_('Unable to convert the value %(value)s'
' to the expected data type %(type)s.') %
{'value': self.value, 'type': type})
raise ClientSideError(msg)
except TypeError:
msg = (_('The data type %(type)s is not supported. The supported'
' data type list is: %(supported)s') %
{'type': type, 'supported': self._supported_types})
raise ClientSideError(msg)
except Exception:
msg = (_('Unexpected exception converting %(value)s to'
' the expected data type %(type)s.') %
{'value': self.value, 'type': type})
raise ClientSideError(msg)
return converted_value
class ProjectNotAuthorized(ClientSideError):
def __init__(self, id, aspect='project'):
params = dict(aspect=aspect, id=id)
super(ProjectNotAuthorized, self).__init__(
_("Not Authorized to access %(aspect)s %(id)s") % params,
def _get_auth_project(on_behalf_of=None):
# when an alarm is created by an admin on behalf of another tenant
# we must ensure for:
# - threshold alarm, that an implicit query constraint on project_id is
# added so that admin-level visibility on statistics is not leaked
# - combination alarm, that alarm ids verification is scoped to
# alarms owned by the alarm project.
# hence for null auth_project (indicating admin-ness) we check if
# the creating tenant differs from the tenant on whose behalf the
# alarm is being created
auth_project = acl.get_limited_to_project(pecan.request.headers)
created_by = pecan.request.headers.get('X-Project-Id')
is_admin = auth_project is None
if is_admin and on_behalf_of != created_by:
auth_project = on_behalf_of
return auth_project
def _sanitize_query(query, db_func, on_behalf_of=None):
"""Check the query.
See if:
1) the request is coming from admin - then allow full visibility
2) non-admin - make sure that the query includes the requester's project.
q = copy.copy(query)
auth_project = _get_auth_project(on_behalf_of)
if auth_project:
_verify_query_segregation(q, auth_project)
proj_q = [i for i in q if i.field == 'project_id']
valid_keys = inspect.getargspec(db_func)[0]
if not proj_q and 'on_behalf_of' not in valid_keys:
# The user is restricted, but they didn't specify a project
# so add it for them.
return q
def _verify_query_segregation(query, auth_project=None):
"""Ensure non-admin queries are not constrained to another project."""
auth_project = (auth_project or
if not auth_project:
for q in query:
if q.field in ('project', 'project_id') and auth_project != q.value:
raise ProjectNotAuthorized(q.value)
def _validate_query(query, db_func, internal_keys=None,
"""Validates the syntax of the query and verifies the query.
Verification check if the query request is authorized for the included
:param query: Query expression that should be validated
:param db_func: the function on the storage level, of which arguments
will form the valid_keys list, which defines the valid fields for a
query expression
:param internal_keys: internally used field names, that should not be
used for querying
:param allow_timestamps: defines whether the timestamp-based constraint is
applicable for this query or not
:returns: None, if the query is valid
:raises InvalidInput: if an operator is not supported for a given field
:raises InvalidInput: if timestamp constraints are allowed, but
search_offset was included without timestamp constraint
:raises: UnknownArgument: if a field name is not a timestamp field, nor
in the list of valid keys
internal_keys = internal_keys or []
valid_keys = inspect.getargspec(db_func)[0]
valid_keys = set(valid_keys) - set(internal_keys)
translation = {'user_id': 'user',
'project_id': 'project',
'resource_id': 'resource'}
has_timestamp_query = _validate_timestamp_fields(query,
('lt', 'le', 'gt', 'ge'),
has_search_offset_query = _validate_timestamp_fields(query,
if has_search_offset_query and not has_timestamp_query:
raise wsme.exc.InvalidInput('field', 'search_offset',
"search_offset cannot be used without " +
def _is_field_metadata(field):
return (field.startswith('metadata.') or
for i in query:
if i.field not in ('timestamp', 'search_offset'):
key = translation.get(i.field, i.field)
operator = i.op
if key in valid_keys or _is_field_metadata(i.field):
if operator == 'eq':
if key == 'enabled':
elif _is_field_metadata(key):
raise wsme.exc.InvalidInput('op', i.op,
'unimplemented operator for '
'%s' % i.field)
msg = ("unrecognized field in query: %s, "
"valid keys: %s") % (query, sorted(valid_keys))
raise wsme.exc.UnknownArgument(key, msg)
def _validate_timestamp_fields(query, field_name, operator_list,
"""Validates the timestamp related constraints in a query if there are any.
:param query: query expression that may contain the timestamp fields
:param field_name: timestamp name, which should be checked (timestamp,
:param operator_list: list of operators that are supported for that
timestamp, which was specified in the parameter field_name
:param allow_timestamps: defines whether the timestamp-based constraint is
applicable to this query or not
:returns: True, if there was a timestamp constraint, containing
a timestamp field named as defined in field_name, in the query and it
was allowed and syntactically correct.
:returns: False, if there wasn't timestamp constraint, containing a
timestamp field named as defined in field_name, in the query
:raises InvalidInput: if an operator is unsupported for a given timestamp
:raises UnknownArgument: if the timestamp constraint is not allowed in
the query
for item in query:
if item.field == field_name:
# If *timestamp* or *search_offset* field was specified in the
# query, but timestamp is not supported on that resource, on
# which the query was invoked, then raise an exception.
if not allow_timestamps:
raise wsme.exc.UnknownArgument(field_name,
"not valid for " +
"this resource")
if item.op not in operator_list:
raise wsme.exc.InvalidInput('op', item.op,
'unimplemented operator for %s' %
return True
return False
def _query_to_kwargs(query, db_func, internal_keys=None,
internal_keys = internal_keys or []
_validate_query(query, db_func, internal_keys=internal_keys,
query = _sanitize_query(query, db_func)
valid_keys = set(inspect.getargspec(db_func)[0]) - set(internal_keys)
translation = {'user_id': 'user',
'project_id': 'project',
'resource_id': 'resource'}
stamp = {}
metaquery = {}
kwargs = {}
for i in query:
if i.field == 'timestamp':
if i.op in ('lt', 'le'):
stamp['end_timestamp'] = i.value
stamp['end_timestamp_op'] = i.op
elif i.op in ('gt', 'ge'):
stamp['start_timestamp'] = i.value
stamp['start_timestamp_op'] = i.op
if i.op == 'eq':
if i.field == 'search_offset':
stamp['search_offset'] = i.value
elif i.field == 'enabled':
kwargs[i.field] = i._get_value_as_type('boolean')
elif i.field.startswith('metadata.'):
metaquery[i.field] = i._get_value_as_type()
elif i.field.startswith('resource_metadata.'):
metaquery[i.field[9:]] = i._get_value_as_type()
key = translation.get(i.field, i.field)
kwargs[key] = i.value
if metaquery and 'metaquery' in valid_keys:
kwargs['metaquery'] = metaquery
if stamp:
q_ts = _get_query_timestamps(stamp)
if 'start' in valid_keys:
kwargs['start'] = q_ts['query_start']
kwargs['end'] = q_ts['query_end']
elif 'start_timestamp' in valid_keys:
kwargs['start_timestamp'] = q_ts['query_start']
kwargs['end_timestamp'] = q_ts['query_end']
if 'start_timestamp_op' in stamp:
kwargs['start_timestamp_op'] = stamp['start_timestamp_op']
if 'end_timestamp_op' in stamp:
kwargs['end_timestamp_op'] = stamp['end_timestamp_op']
return kwargs
def _validate_groupby_fields(groupby_fields):
"""Checks that the list of groupby fields from request is valid.
If all fields are valid, returns fields with duplicates removed.
# NOTE(terriyu): Currently, metadata fields are supported in our
# group by statistics implementation only for mongodb
valid_fields = set(['user_id', 'resource_id', 'project_id', 'source',
invalid_fields = set(groupby_fields) - valid_fields
if invalid_fields:
raise wsme.exc.UnknownArgument(invalid_fields,
"Invalid groupby fields")
# Remove duplicate fields
# NOTE(terriyu): This assumes that we don't care about the order of the
# group by fields.
return list(set(groupby_fields))
def _get_query_timestamps(args=None):
"""Return any optional timestamp information in the request.
Determine the desired range, if any, from the GET arguments. Set
up the query range using the specified offset.
[query_start ... start_timestamp ... end_timestamp ... query_end]
Returns a dictionary containing:
query_start: First timestamp to use for query
start_timestamp: start_timestamp parameter from request
query_end: Final timestamp to use for query
end_timestamp: end_timestamp parameter from request
search_offset: search_offset parameter from request
if args is None:
return {'query_start': None,
'query_end': None,
'start_timestamp': None,
'end_timestamp': None,
'search_offset': 0}
search_offset = int(args.get('search_offset', 0))
def _parse_timestamp(timestamp):
if not timestamp:
return None
iso_timestamp = timeutils.parse_isotime(timestamp)
iso_timestamp = iso_timestamp.replace(tzinfo=None)
except ValueError:
raise wsme.exc.InvalidInput('timestamp', timestamp,
'invalid timestamp format')
return iso_timestamp
start_timestamp = args.get('start_timestamp')
end_timestamp = args.get('end_timestamp')
start_timestamp = _parse_timestamp(start_timestamp)
end_timestamp = _parse_timestamp(end_timestamp)
query_start = start_timestamp - datetime.timedelta(
minutes=search_offset) if start_timestamp else None
query_end = end_timestamp + datetime.timedelta(
minutes=search_offset) if end_timestamp else None
return {'query_start': query_start,
'query_end': query_end,
'start_timestamp': start_timestamp,
'end_timestamp': end_timestamp,
'search_offset': search_offset,
def _flatten_metadata(metadata):
"""Return flattened resource metadata.
Metadata is returned with flattened nested structures (except nested sets)
and with all values converted to unicode strings.
if metadata:
# After changing recursive_keypairs` output we need to keep
# flattening output unchanged.
# Example: recursive_keypairs({'a': {'b':{'c':'d'}}}, '.')
# output before: a.b:c=d
# output now: a.b.c=d
# So to keep the first variant just replace all dots except the first
return dict((k.replace('.', ':').replace(':', '.', 1),
for k, v in utils.recursive_keypairs(metadata,
if type(v) is not set)
return {}
def _make_link(rel_name, url, type, type_arg, query=None):
query_str = ''
if query:
query_str = '?q.field=%s&q.value=%s' % (query['field'],
return Link(href='%s/v2/%s/%s%s' % (url, type, type_arg, query_str),
def _send_notification(event, payload):
notification = event.replace(" ", "_")
notification = "alarm.%s" % notification
transport = messaging.get_transport()
notifier = messaging.get_notifier(transport, publisher_id="ceilometer.api")
# FIXME(sileht): perhaps we need to copy some infos from the
# pecan request headers like nova does, notification, payload)
class OldSample(_Base):
"""A single measurement for a given meter and resource.
This class is deprecated in favor of Sample.
source = wtypes.text
"The ID of the source that identifies where the sample comes from"
counter_name = wsme.wsattr(wtypes.text, mandatory=True)
"The name of the meter"
# FIXME(dhellmann): Make this meter_name?
counter_type = wsme.wsattr(wtypes.text, mandatory=True)
"The type of the meter (see :ref:`measurements`)"
# FIXME(dhellmann): Make this meter_type?
counter_unit = wsme.wsattr(wtypes.text, mandatory=True)
"The unit of measure for the value in counter_volume"
# FIXME(dhellmann): Make this meter_unit?
counter_volume = wsme.wsattr(float, mandatory=True)
"The actual measured value"
user_id = wtypes.text
"The ID of the user who last triggered an update to the resource"
project_id = wtypes.text
"The ID of the project or tenant that owns the resource"
resource_id = wsme.wsattr(wtypes.text, mandatory=True)
"The ID of the :class:`Resource` for which the measurements are taken"
timestamp = datetime.datetime
"UTC date and time when the measurement was made"
recorded_at = datetime.datetime
"When the sample has been recorded."
resource_metadata = {wtypes.text: wtypes.text}
"Arbitrary metadata associated with the resource"
message_id = wtypes.text
"A unique identifier for the sample"
def __init__(self, counter_volume=None, resource_metadata=None,
timestamp=None, **kwds):
resource_metadata = resource_metadata or {}
if counter_volume is not None:
counter_volume = float(counter_volume)
resource_metadata = _flatten_metadata(resource_metadata)
# this is to make it easier for clients to pass a timestamp in
if timestamp and isinstance(timestamp, six.string_types):
timestamp = timeutils.parse_isotime(timestamp)
super(OldSample, self).__init__(counter_volume=counter_volume,
timestamp=timestamp, **kwds)
if self.resource_metadata in (wtypes.Unset, None):
self.resource_metadata = {}
def sample(cls):
return cls(source='openstack',
resource_metadata={'name1': 'value1',
'name2': 'value2'},
class Statistics(_Base):
"""Computed statistics for a query."""
groupby = {wtypes.text: wtypes.text}
"Dictionary of field names for group, if groupby statistics are requested"
unit = wtypes.text
"The unit type of the data set"
min = float
"The minimum volume seen in the data"
max = float
"The maximum volume seen in the data"
avg = float
"The average of all of the volume values seen in the data"
sum = float
"The total of all of the volume values seen in the data"
count = int
"The number of samples seen"
aggregate = {wtypes.text: float}
"The selectable aggregate value(s)"
duration = float
"The difference, in seconds, between the oldest and newest timestamp"
duration_start = datetime.datetime
"UTC date and time of the earliest timestamp, or the query start time"
duration_end = datetime.datetime
"UTC date and time of the oldest timestamp, or the query end time"
period = int
"The difference, in seconds, between the period start and end"
period_start = datetime.datetime
"UTC date and time of the period start"
period_end = datetime.datetime
"UTC date and time of the period end"
def __init__(self, start_timestamp=None, end_timestamp=None, **kwds):
super(Statistics, self).__init__(**kwds)
self._update_duration(start_timestamp, end_timestamp)
def _update_duration(self, start_timestamp, end_timestamp):
# "Clamp" the timestamps we return to the original time
# range, excluding the offset.
if (start_timestamp and
self.duration_start and
self.duration_start < start_timestamp):
self.duration_start = start_timestamp
LOG.debug(_('clamping min timestamp to range'))
if (end_timestamp and
self.duration_end and
self.duration_end > end_timestamp):
self.duration_end = end_timestamp
LOG.debug(_('clamping max timestamp to range'))
# If we got valid timestamps back, compute a duration in seconds.
# If the min > max after clamping then we know the
# timestamps on the samples fell outside of the time
# range we care about for the query, so treat them as
# "invalid."
# If the timestamps are invalid, return None as a
# sentinel indicating that there is something "funny"
# about the range.
if (self.duration_start and
self.duration_end and
self.duration_start <= self.duration_end):
self.duration = timeutils.delta_seconds(self.duration_start,
self.duration_start = self.duration_end = self.duration = None
def sample(cls):
return cls(unit='GiB',
duration_start=datetime.datetime(2013, 1, 4, 16, 42),
duration_end=datetime.datetime(2013, 1, 4, 16, 47),
period_start=datetime.datetime(2013, 1, 4, 16, 00),
period_end=datetime.datetime(2013, 1, 4, 18, 00),
class Aggregate(_Base):
func = wsme.wsattr(wtypes.text, mandatory=True)
"The aggregation function name"
param = wsme.wsattr(wtypes.text, default=None)
"The paramter to the aggregation function"
def __init__(self, **kwargs):
super(Aggregate, self).__init__(**kwargs)
def validate(aggregate):
return aggregate
def sample(cls):
return cls(func='cardinality',
class MeterController(rest.RestController):
"""Manages operations on a single meter."""
_custom_actions = {
'statistics': ['GET'],
def __init__(self, meter_name):
pecan.request.context['meter_name'] = meter_name
self.meter_name = meter_name
@wsme_pecan.wsexpose([OldSample], [Query], int)
def get_all(self, q=None, limit=None):
"""Return samples for the meter.
:param q: Filter rules for the data to be returned.
:param limit: Maximum number of samples to return.
q = q or []
if limit and limit < 0:
raise ClientSideError(_("Limit must be positive"))
kwargs = _query_to_kwargs(q, storage.SampleFilter.__init__)
kwargs['meter'] = self.meter_name
f = storage.SampleFilter(**kwargs)
return [OldSample.from_db_model(e)
for e in pecan.request.storage_conn.get_samples(f, limit=limit)
@wsme_pecan.wsexpose([OldSample], body=[OldSample])
def post(self, samples):
"""Post a list of new Samples to Telemetry.
:param samples: a list of samples within the request body.
now = timeutils.utcnow()
auth_project = acl.get_limited_to_project(pecan.request.headers)
def_source = pecan.request.cfg.sample_source
def_project_id = pecan.request.headers.get('X-Project-Id')
def_user_id = pecan.request.headers.get('X-User-Id')
published_samples = []
for s in samples:
if self.meter_name != s.counter_name:
raise wsme.exc.InvalidInput('counter_name', s.counter_name,
'should be %s' % self.meter_name)
if s.message_id:
raise wsme.exc.InvalidInput('message_id', s.message_id,
'The message_id must not be set')
if s.counter_type not in sample.TYPES:
raise wsme.exc.InvalidInput('counter_type', s.counter_type,
'The counter type must be: ' +
', '.join(sample.TYPES))
s.user_id = (s.user_id or def_user_id)
s.project_id = (s.project_id or def_project_id)
s.source = '%s:%s' % (s.project_id, (s.source or def_source))
s.timestamp = (s.timestamp or now)
if auth_project and auth_project != s.project_id:
# non admin user trying to cross post to another project_id
auth_msg = 'can not post samples to other projects'
raise wsme.exc.InvalidInput('project_id', s.project_id,
published_sample = sample.Sample(
s.message_id =
with pecan.request.pipeline_manager.publisher(
context.get_admin_context()) as publisher:
return samples
@wsme_pecan.wsexpose([Statistics], [Query], [unicode], int, [Aggregate])
def statistics(self, q=None, groupby=None, period=None, aggregate=None):
"""Computes the statistics of the samples in the time range given.
:param q: Filter rules for the data to be returned.
:param groupby: Fields for group by aggregation
:param period: Returned result will be an array of statistics for a
period long of that number of seconds.
:param aggregate: The selectable aggregation functions to be applied.
q = q or []
groupby = groupby or []
aggregate = aggregate or []
if period and period < 0:
raise ClientSideError(_("Period must be positive."))
kwargs = _query_to_kwargs(q, storage.SampleFilter.__init__)
kwargs['meter'] = self.meter_name
f = storage.SampleFilter(**kwargs)
g = _validate_groupby_fields(groupby)
aggregate = utils.uniq(aggregate, ['func', 'param'])
computed = pecan.request.storage_conn.get_meter_statistics(f,
LOG.debug(_('computed value coming from %r'),
# Find the original timestamp in the query to use for clamping
# the duration returned in the statistics.
start = end = None
for i in q:
if i.field == 'timestamp' and i.op in ('lt', 'le'):
end = timeutils.parse_isotime(i.value).replace(tzinfo=None)
elif i.field == 'timestamp' and i.op in ('gt', 'ge'):
start = timeutils.parse_isotime(i.value).replace(tzinfo=None)
return [Statistics(start_timestamp=start,
for c in computed]
class Meter(_Base):
"""One category of measurements."""
name = wtypes.text
"The unique name for the meter"
type = wtypes.Enum(str, *sample.TYPES)
"The meter type (see :ref:`measurements`)"
unit = wtypes.text
"The unit of measure"
resource_id = wtypes.text
"The ID of the :class:`Resource` for which the measurements are taken"
project_id = wtypes.text
"The ID of the project or tenant that owns the resource"
user_id = wtypes.text
"The ID of the user who last triggered an update to the resource"
source = wtypes.text
"The ID of the source that identifies where the meter comes from"
meter_id = wtypes.text
"The unique identifier for the meter"
def __init__(self, **kwargs):
meter_id = '%s+%s' % (kwargs['resource_id'], kwargs['name'])
# meter_id is of type Unicode but base64.encodestring() only accepts
# strings. See bug #1333177
meter_id = base64.encodestring(meter_id.encode('utf-8'))
kwargs['meter_id'] = meter_id
super(Meter, self).__init__(**kwargs)
def sample(cls):
return cls(name='instance',
class MetersController(rest.RestController):
"""Works on meters."""
def _lookup(self, meter_name, *remainder):
return MeterController(meter_name), remainder
@wsme_pecan.wsexpose([Meter], [Query])
def get_all(self, q=None):
"""Return all known meters, based on the data recorded so far.
:param q: Filter rules for the meters to be returned.
q = q or []
# Timestamp field is not supported for Meter queries
kwargs = _query_to_kwargs(q, pecan.request.storage_conn.get_meters,
return [Meter.from_db_model(m)
for m in pecan.request.storage_conn.get_meters(**kwargs)]
class Sample(_Base):
"""One measurement."""
id = wtypes.text
"The unique identifier for the sample."
meter = wtypes.text
"The meter name this sample is for."
type = wtypes.Enum(str, *sample.TYPES)
"The meter type (see :ref:`measurements`)"
unit = wtypes.text
"The unit of measure."
volume = float
"The metered value."
user_id = wtypes.text
"The user this sample was taken for."
project_id = wtypes.text
"The project this sample was taken for."
resource_id = wtypes.text
"The :class:`Resource` this sample was taken for."
source = wtypes.text
"The source that identifies where the sample comes from."
timestamp = datetime.datetime
"When the sample has been generated."
recorded_at = datetime.datetime
"When the sample has been recorded."
metadata = {wtypes.text: wtypes.text}
"Arbitrary metadata associated with the sample."
def from_db_model(cls, m):
return cls(id=m.message_id,
def sample(cls):
return cls(id=str(uuid.uuid1()),
metadata={'name1': 'value1',
'name2': 'value2'},
class SamplesController(rest.RestController):
"""Controller managing the samples."""
@wsme_pecan.wsexpose([Sample], [Query], int)
def get_all(self, q=None, limit=None):
"""Return all known samples, based on the data recorded so far.
:param q: Filter rules for the samples to be returned.
:param limit: Maximum number of samples to be returned.
q = q or []
if limit and limit < 0:
raise ClientSideError(_("Limit must be positive"))
kwargs = _query_to_kwargs(q, storage.SampleFilter.__init__)
f = storage.SampleFilter(**kwargs)
return map(Sample.from_db_model,
pecan.request.storage_conn.get_samples(f, limit=limit))
@wsme_pecan.wsexpose(Sample, wtypes.text)
def get_one(self, sample_id):
"""Return a sample.
:param sample_id: the id of the sample.
f = storage.SampleFilter(message_id=sample_id)
samples = list(pecan.request.storage_conn.get_samples(f))
if len(samples) < 1:
raise EntityNotFound(_('Sample'), sample_id)
return Sample.from_db_model(samples[0])
class ComplexQuery(_Base):
"""Holds a sample query encoded in json."""
filter = wtypes.text
"The filter expression encoded in json."
orderby = wtypes.text
"List of single-element dicts for specifing the ordering of the results."
limit = int
"The maximum number of results to be returned."
def sample(cls):
return cls(filter='{"and": [{"and": [{"=": ' +
'{"counter_name": "cpu_util"}}, ' +
'{">": {"counter_volume": 0.23}}, ' +
'{"<": {"counter_volume": 0.26}}]}, ' +
'{"or": [{"and": [{">": ' +
'{"timestamp": "2013-12-01T18:00:00"}}, ' +
'{"<": ' +
'{"timestamp": "2013-12-01T18:15:00"}}]}, ' +
'{"and": [{">": ' +
'{"timestamp": "2013-12-01T18:30:00"}}, ' +
'{"<": ' +
'{"timestamp": "2013-12-01T18:45:00"}}]}]}]}',
orderby='[{"counter_volume": "ASC"}, ' +
'{"timestamp": "DESC"}]',
def _list_to_regexp(items, regexp_prefix=""):
regexp = ["^%s$" % item for item in items]
regexp = regexp_prefix + "|".join(regexp)
return regexp
class ValidatedComplexQuery(object):
complex_operators = ["and", "or"]
order_directions = ["asc", "desc"]
simple_ops = ["=", "!=", "<", ">", "<=", "=<", ">=", "=>"]
regexp_prefix = "(?i)"
complex_ops = _list_to_regexp(complex_operators, regexp_prefix)
simple_ops = _list_to_regexp(simple_ops, regexp_prefix)
order_directions = _list_to_regexp(order_directions, regexp_prefix)
timestamp_fields = ["timestamp", "state_timestamp"]
def __init__(self, query, db_model, additional_name_mapping=None,
additional_name_mapping = additional_name_mapping or {}
self.name_mapping = {"user": "user_id",
"project": "project_id"}
valid_keys = db_model.get_field_names()
valid_keys = list(valid_keys) + self.name_mapping.keys()
valid_fields = _list_to_regexp(valid_keys)
if metadata_allowed:
valid_filter_fields = valid_fields + "|^metadata\.[\S]+$"
valid_filter_fields = valid_fields
schema_value = {
"oneOf": [{"type": "string"},
{"type": "number"},
{"type": "boolean"}],
"minProperties": 1,
"maxProperties": 1}
schema_value_in = {
"type": "array",
"items": {"oneOf": [{"type": "string"},
{"type": "number"}]},
"minItems": 1}
schema_field = {
"type": "object",
"patternProperties": {valid_filter_fields: schema_value},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}
schema_field_in = {
"type": "object",
"patternProperties": {valid_filter_fields: schema_value_in},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}
schema_leaf_in = {
"type": "object",
"patternProperties": {"(?i)^in$": schema_field_in},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}
schema_leaf_simple_ops = {
"type": "object",
"patternProperties": {self.simple_ops: schema_field},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}
schema_and_or_array = {
"type": "array",
"items": {"$ref": "#"},
"minItems": 2}
schema_and_or = {
"type": "object",
"patternProperties": {self.complex_ops: schema_and_or_array},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}
schema_not = {
"type": "object",
"patternProperties": {"(?i)^not$": {"$ref": "#"}},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}
self.schema = {
"oneOf": [{"$ref": "#/definitions/leaf_simple_ops"},
{"$ref": "#/definitions/leaf_in"},
{"$ref": "#/definitions/and_or"},
{"$ref": "#/definitions/not"}],
"minProperties": 1,
"maxProperties": 1,
"definitions": {"leaf_simple_ops": schema_leaf_simple_ops,
"leaf_in": schema_leaf_in,
"and_or": schema_and_or,
"not": schema_not}}
self.orderby_schema = {
"type": "array",
"items": {
"type": "object",
{"type": "string",
"pattern": self.order_directions}},
"additionalProperties": False,
"minProperties": 1,
"maxProperties": 1}}
self.original_query = query
def validate(self, visibility_field):
"""Validates the query content and does the necessary conversions."""
if self.original_query.filter is wtypes.Unset:
self.filter_expr = None
self.filter_expr = json.loads(self.original_query.filter)
except (ValueError, jsonschema.exceptions.ValidationError) as e:
raise ClientSideError(_("Filter expression not valid: %s") %
if self.original_query.orderby is wtypes.Unset:
self.orderby = None
self.orderby = json.loads(self.original_query.orderby)
except (ValueError, jsonschema.exceptions.ValidationError) as e:
raise ClientSideError(_("Order-by expression not valid: %s") %
if self.original_query.limit is wtypes.Unset:
self.limit = None
self.limit = self.original_query.limit
if self.limit is not None and self.limit <= 0:
msg = _('Limit should be positive')
raise ClientSideError(msg)
def _convert_orderby_to_lower_case(orderby):
for orderby_field in orderby:
def _normalize_field_names_in_orderby(self, orderby):
for orderby_field in orderby:
def _traverse_postorder(self, tree, visitor):
op = tree.keys()[0]
if op.lower() in self.complex_operators:
for i, operand in enumerate(tree[op]):
self._traverse_postorder(operand, visitor)
if op.lower() == "not":
self._traverse_postorder(tree[op], visitor)
def _check_cross_project_references(self, own_project_id,
"""Do not allow other than own_project_id."""
def check_project_id(subfilter):
op = subfilter.keys()[0]
if (op.lower() not in self.complex_operators
and subfilter[op].keys()[0] == visibility_field
and subfilter[op][visibility_field] != own_project_id):
raise ProjectNotAuthorized(subfilter[op][visibility_field])
self._traverse_postorder(self.filter_expr, check_project_id)
def _force_visibility(self, visibility_field):
"""Force visibility field.
If the tenant is not admin insert an extra
"and <visibility_field>=<tenant's project_id>" clause to the query.
authorized_project = acl.get_limited_to_project(pecan.request.headers)
is_admin = authorized_project is None
if not is_admin:
self._restrict_to_project(authorized_project, visibility_field)
def _restrict_to_project(self, project_id, visibility_field):
restriction = {"=": {visibility_field: project_id}}
if self.filter_expr is None:
self.filter_expr = restriction
self.filter_expr = {"and": [restriction, self.filter_expr]}
def _replace_isotime_with_datetime(self, filter_expr):
def replace_isotime(subfilter):
op = subfilter.keys()[0]
if (op.lower() not in self.complex_operators
and subfilter[op].keys()[0] in self.timestamp_fields):
field = subfilter[op].keys()[0]
date_time = self._convert_to_datetime(subfilter[op][field])
subfilter[op][field] = date_time
self._traverse_postorder(filter_expr, replace_isotime)
def _normalize_field_names_for_db_model(self, filter_expr):
def _normalize_field_names(subfilter):
op = subfilter.keys()[0]
if op.lower() not in self.complex_operators:
def _replace_field_names(self, subfilter):
field = subfilter.keys()[0]
value = subfilter[field]
if field in self.name_mapping:
del subfilter[field]
subfilter[self.name_mapping[field]] = value
if field.startswith("metadata."):
del subfilter[field]
subfilter["resource_" + field] = value
def _convert_operator_to_lower_case(self, filter_expr):
self._traverse_postorder(filter_expr, utils.lowercase_keys)
def _convert_to_datetime(isotime):
date_time = timeutils.parse_isotime(isotime)
date_time = date_time.replace(tzinfo=None)
return date_time
except ValueError:
LOG.exception(_("String %s is not a valid isotime") % isotime)
msg = _('Failed to parse the timestamp value %s') % isotime
raise ClientSideError(msg)
def _validate_filter(self, filter_expr):
jsonschema.validate(filter_expr, self.schema)
def _validate_orderby(self, orderby_expr):
jsonschema.validate(orderby_expr, self.orderby_schema)
class Resource(_Base):
"""An externally defined object for which samples have been received."""
resource_id = wtypes.text
"The unique identifier for the resource"
project_id = wtypes.text
"The ID of the owning project or tenant"
user_id = wtypes.text
"The ID of the user who created the resource or updated it last"
first_sample_timestamp = datetime.datetime
"UTC date & time not later than the first sample known for this resource"
last_sample_timestamp = datetime.datetime
"UTC date & time not earlier than the last sample known for this resource"
metadata = {wtypes.text: wtypes.text}
"Arbitrary metadata associated with the resource"
links = [Link]
"A list containing a self link and associated meter links"
source = wtypes.text
"The source where the resource come from"
def __init__(self, metadata=None, **kwds):
metadata = metadata or {}
metadata = _flatten_metadata(metadata)
super(Resource, self).__init__(metadata=metadata, **kwds)
def sample(cls):
return cls(resource_id='bd9431c1-8d69-4ad3-803a-8d4a6b89fd36',
metadata={'name1': 'value1',
'name2': 'value2'},
class ResourcesController(rest.RestController):
"""Works on resources."""
def _resource_links(self, resource_id, meter_links=1):
links = [_make_link('self', pecan.request.host_url, 'resources',
if meter_links:
for meter in pecan.request.storage_conn.get_meters(
query = {'field': 'resource_id', 'value': resource_id}
links.append(_make_link(, pecan.request.host_url,
'meters',, query=query))
return links
@wsme_pecan.wsexpose(Resource, unicode)
def get_one(self, resource_id):
"""Retrieve details about one resource.
:param resource_id: The UUID of the resource.
authorized_project = acl.get_limited_to_project(pecan.request.headers)
resources = list(pecan.request.storage_conn.get_resources(
resource=resource_id, project=authorized_project))
if not resources:
raise EntityNotFound(_('Resource'), resource_id)
return Resource.from_db_and_links(resources[0],
@wsme_pecan.wsexpose([Resource], [Query], int)
def get_all(self, q=None, meter_links=1):
"""Retrieve definitions of all of the resources.
:param q: Filter rules for the resources to be returned.
:param meter_links: option to include related meter links
q = q or []
kwargs = _query_to_kwargs(q, pecan.request.storage_conn.get_resources)
resources = [
for r in pecan.request.storage_conn.get_resources(**kwargs)]
return resources
class AlarmThresholdRule(_Base):
meter_name = wsme.wsattr(wtypes.text, mandatory=True)
"The name of the meter"
# FIXME(sileht): default doesn't work
# workaround: default is set in validate method
query = wsme.wsattr([Query], default=[])
"""The query to find the data for computing statistics.
Ownership settings are automatically included based on the Alarm owner.
period = wsme.wsattr(wtypes.IntegerType(minimum=1), default=60)
"The time range in seconds over which query"
comparison_operator = AdvEnum('comparison_operator', str,
'lt', 'le', 'eq', 'ne', 'ge', 'gt',
"The comparison against the alarm threshold"
threshold = wsme.wsattr(float, mandatory=True)
"The threshold of the alarm"
statistic = AdvEnum('statistic', str, 'max', 'min', 'avg', 'sum',
'count', default='avg')
"The statistic to compare to the threshold"
evaluation_periods = wsme.wsattr(wtypes.IntegerType(minimum=1), default=1)
"The number of historical periods to evaluate the threshold"
exclude_outliers = wsme.wsattr(bool, default=False)
"Whether datapoints with anomalously low sample counts are excluded"
def __init__(self, query=None, **kwargs):
if query:
query = [Query(**q) for q in query]
super(AlarmThresholdRule, self).__init__(query=query, **kwargs)
def validate(threshold_rule):
# note(sileht): wsme default doesn't work in some case
# workaround for
if not threshold_rule.query:
threshold_rule.query = []
# Timestamp is not allowed for AlarmThresholdRule query, as the alarm
# evaluator will construct timestamp bounds for the sequence of
# statistics queries as the sliding evaluation window advances
# over time.
_validate_query(threshold_rule.query, storage.SampleFilter.__init__,
return threshold_rule
def default_description(self):
return (_('Alarm when %(meter_name)s is %(comparison_operator)s a '
'%(statistic)s of %(threshold)s over %(period)s seconds') %
def as_dict(self):
rule = self.as_dict_from_keys(['period', 'comparison_operator',
'threshold', 'statistic',
'evaluation_periods', 'meter_name',
rule['query'] = [q.as_dict() for q in self.query]
return rule
def sample(cls):
return cls(meter_name='cpu_util',
query=[{'field': 'resource_id',
'value': '2a4d689b-f0b8-49c1-9eef-87cae58d80db',
'op': 'eq',
'type': 'string'}])
class AlarmCombinationRule(_Base):
operator = AdvEnum('operator', str, 'or', 'and', default='and')
"How to combine the sub-alarms"
alarm_ids = wsme.wsattr([wtypes.text], mandatory=True)
"List of alarm identifiers to combine"
def default_description(self):
joiner = ' %s ' % self.operator
return _('Combined state of alarms %s') % joiner.join(self.alarm_ids)
def as_dict(self):
return self.as_dict_from_keys(['operator', 'alarm_ids'])
def validate(rule):
rule.alarm_ids = sorted(set(rule.alarm_ids), key=rule.alarm_ids.index)
if len(rule.alarm_ids) <= 1:
raise ClientSideError(_('Alarm combination rule should contain at'
' least two different alarm ids.'))
return rule
def sample(cls):
return cls(operator='or',
class AlarmTimeConstraint(_Base):
"""Representation of a time constraint on an alarm."""
name = wsme.wsattr(wtypes.text, mandatory=True)
"The name of the constraint"
_description = None # provide a default
def get_description(self):
if not self._description:
return ('Time constraint at %s lasting for %s seconds'
% (self.start, self.duration))
return self._description
def set_description(self, value):
self._description = value
description = wsme.wsproperty(wtypes.text, get_description,
"The description of the constraint"
start = wsme.wsattr(CronType(), mandatory=True)
"Start point of the time constraint, in cron format"
duration = wsme.wsattr(wtypes.IntegerType(minimum=0), mandatory=True)
"How long the constraint should last, in seconds"
timezone = wsme.wsattr(wtypes.text, default="")
"Timezone of the constraint"
def as_dict(self):
return self.as_dict_from_keys(['name', 'description', 'start',
'duration', 'timezone'])
def validate(tc):
if tc.timezone:
except Exception:
raise ClientSideError(_("Timezone %s is not valid")
% tc.timezone)
return tc
def sample(cls):
return cls(name='SampleConstraint',
description='nightly build every night at 23h for 3 hours',
start='0 23 * * *',
class Alarm(_Base):
"""Representation of an alarm.
.. note::
combination_rule and threshold_rule are mutually exclusive. The *type*
of the alarm should be set to *threshold* or *combination* and the
appropriate rule should be filled.
alarm_id = wtypes.text
"The UUID of the alarm"
name = wsme.wsattr(wtypes.text, mandatory=True)
"The name for the alarm"
_description = None # provide a default
def get_description(self):
rule = getattr(self, '%s_rule' % self.type, None)
if not self._description and rule:
return six.text_type(rule.default_description)
return self._description
def set_description(self, value):
self._description = value
description = wsme.wsproperty(wtypes.text, get_description,
"The description of the alarm"
enabled = wsme.wsattr(bool, default=True)
"This alarm is enabled?"
ok_actions = wsme.wsattr([wtypes.text], default=[])
"The actions to do when alarm state change to ok"
alarm_actions = wsme.wsattr([wtypes.text], default=[])
"The actions to do when alarm state change to alarm"
insufficient_data_actions = wsme.wsattr([wtypes.text], default=[])
"The actions to do when alarm state change to insufficient data"
repeat_actions = wsme.wsattr(bool, default=False)
"The actions should be re-triggered on each evaluation cycle"
type = AdvEnum('type', str, 'threshold', 'combination', mandatory=True)
"Explicit type specifier to select which rule to follow below."
threshold_rule = AlarmThresholdRule
"Describe when to trigger the alarm based on computed statistics"
combination_rule = AlarmCombinationRule
"""Describe when to trigger the alarm based on combining the state of
other alarms"""
time_constraints = wtypes.wsattr([AlarmTimeConstraint], default=[])
"""Describe time constraints for the alarm"""
# These settings are ignored in the PUT or POST operations, but are
# filled in for GET
project_id = wtypes.text
"The ID of the project or tenant that owns the alarm"
user_id = wtypes.text
"The ID of the user who created the alarm"
timestamp = datetime.datetime
"The date of the last alarm definition update"
state = AdvEnum('state', str, *state_kind,
default='insufficient data')
"The state offset the alarm"
state_timestamp = datetime.datetime
"The date of the last alarm state changed"
def __init__(self, rule=None, time_constraints=None, **kwargs):
super(Alarm, self).__init__(**kwargs)
if rule:
if self.type == 'threshold':
self.threshold_rule = AlarmThresholdRule(**rule)
elif self.type == 'combination':
self.combination_rule = AlarmCombinationRule(**rule)
if time_constraints:
self.time_constraints = [AlarmTimeConstraint(**tc)
for tc in time_constraints]
def validate(alarm):
if alarm.threshold_rule:
# ensure an implicit constraint on project_id is added to
# the query if not already present
alarm.threshold_rule.query = _sanitize_query(
elif alarm.combination_rule:
project = _get_auth_project(alarm.project_id
if alarm.project_id != wtypes.Unset
else None)
for id in alarm.combination_rule.alarm_ids:
alarms = list(pecan.request.alarm_storage_conn.get_alarms(
alarm_id=id, project=project))
if not alarms:
raise AlarmNotFound(id, project)
tc_names = [ for tc in alarm.time_constraints]
if len(tc_names) > len(set(tc_names)):
error = _("Time constraint names must be "
"unique for a given alarm.")
raise ClientSideError(error)
return alarm
def check_rule(alarm):
rule = '%s_rule' % alarm.type
if getattr(alarm, rule) in (wtypes.Unset, None):
error = _("%(rule)s must be set for %(type)s"
" type alarm") % {"rule": rule, "type": alarm.type}
raise ClientSideError(error)
if alarm.threshold_rule and alarm.combination_rule:
error = _("threshold_rule and combination_rule "
"cannot be set at the same time")
raise ClientSideError(error)
def check_alarm_actions(alarm):
actions_schema = alarm_service.AlarmNotifierService.notifiers_schemas
for state in state_kind:
actions_name = state.replace(" ", "_") + '_actions'
actions = getattr(alarm, actions_name)
if not actions:
for action in actions:
url = netutils.urlsplit(action)
except Exception:
error = _("Unable to parse action %s") % action
raise ClientSideError(error)
if url.scheme not in actions_schema:
error = _("Unsupported action %s") % action
raise ClientSideError(error)
def sample(cls):
return cls(alarm_id=None,
description="An alarm",
def as_dict(self, db_model):
d = super(Alarm, self).as_dict(db_model)
for k in d:
if k.endswith('_rule'):
del d[k]
d['rule'] = getattr(self, "%s_rule" % self.type).as_dict()
d['time_constraints'] = [tc.as_dict() for tc in self.time_constraints]
return d
class AlarmChange(_Base):
"""Representation of an event in an alarm's history."""
event_id = wtypes.text
"The UUID of the change event"
alarm_id = wtypes.text
"The UUID of the alarm"
type = wtypes.Enum(str,
'rule change',
'state transition',
"The type of change"
detail = wtypes.text
"JSON fragment describing change"
project_id = wtypes.text
"The project ID of the initiating identity"
user_id = wtypes.text
"The user ID of the initiating identity"
on_behalf_of = wtypes.text
"The tenant on behalf of which the change is being made"
timestamp = datetime.datetime
"The time/date of the alarm change"
def sample(cls):
return cls(alarm_id='e8ff32f772a44a478182c3fe1f7cad6a',
type='rule change',
detail='{"threshold": 42.0, "evaluation_periods": 4}',
class AlarmController(rest.RestController):
"""Manages operations on a single alarm."""
_custom_actions = {
'history': ['GET'],
'state': ['PUT', 'GET'],
def __init__(self, alarm_id):
pecan.request.context['alarm_id'] = alarm_id
self._id = alarm_id
def _alarm(self):
self.conn = pecan.request.alarm_storage_conn
auth_project = acl.get_limited_to_project(pecan.request.headers)
alarms = list(self.conn.get_alarms(alarm_id=self._id,
if not alarms:
raise AlarmNotFound(alarm=self._id, auth_project=auth_project)
return alarms[0]
def _record_change(self, data, now, on_behalf_of=None, type=None):
if not cfg.CONF.alarm.record_history:
type = type or alarm_models.AlarmChange.RULE_CHANGE
scrubbed_data = utils.stringify_timestamps(data)
detail = json.dumps(scrubbed_data)
user_id = pecan.request.headers.get('X-User-Id')
project_id = pecan.request.headers.get('X-Project-Id')
on_behalf_of = on_behalf_of or project_id
payload = dict(event_id=str(uuid.uuid4()),
except ceilometer.NotImplementedError:
# Revert to the pre-json'ed details ...
payload['detail'] = scrubbed_data
_send_notification(type, payload)
def get(self):
"""Return this alarm."""
return Alarm.from_db_model(self._alarm())
@wsme_pecan.wsexpose(Alarm, body=Alarm)
def put(self, data):
"""Modify this alarm.
:param data: an alarm within the request body.
# Ensure alarm exists
alarm_in = self._alarm()
now = timeutils.utcnow()
data.alarm_id = self._id
user, project = acl.get_limited_to(pecan.request.headers)
if user:
data.user_id = user
elif data.user_id == wtypes.Unset:
data.user_id = alarm_in.user_id
if project:
data.project_id = project
elif data.project_id == wtypes.Unset:
data.project_id = alarm_in.project_id
data.timestamp = now
if alarm_in.state != data.state:
data.state_timestamp = now
data.state_timestamp = alarm_in.state_timestamp
# make sure alarms are unique by name per project.
if !=
alarms = list(self.conn.get_alarms(,
if alarms:
raise ClientSideError(
_("Alarm with name=%s exists") %,
# should check if there is any circle in the dependency, but for
# efficiency reason, here only check alarm cannot depend on itself
if data.type == 'combination':
if self._id in data.combination_rule.alarm_ids:
raise ClientSideError(_('Cannot specify alarm %s itself in '
'combination rule') % self._id)
old_alarm = Alarm.from_db_model(alarm_in).as_dict(alarm_models.Alarm)
updated_alarm = data.as_dict(alarm_models.Alarm)
alarm_in = alarm_models.Alarm(**updated_alarm)
except Exception:
LOG.exception(_("Error while putting alarm: %s") % updated_alarm)
raise ClientSideError(_("Alarm incorrect"))
alarm = self.conn.update_alarm(alarm_in)
change = dict((k, v) for k, v in updated_alarm.items()
if v != old_alarm[k] and k not in
['timestamp', 'state_timestamp'])
self._record_change(change, now, on_behalf_of=alarm.project_id)
return Alarm.from_db_model(alarm)
@wsme_pecan.wsexpose(None, status_code=204)
def delete(self):
"""Delete this alarm."""
# ensure alarm exists before deleting
alarm = self._alarm()
change = Alarm.from_db_model(alarm).as_dict(alarm_models.Alarm)
# TODO(eglynn): add pagination marker to signature once overall
# API support for pagination is finalized
@wsme_pecan.wsexpose([AlarmChange], [Query])
def history(self, q=None):
"""Assembles the alarm history requested.
:param q: Filter rules for the changes to be described.
q = q or []
# allow history to be returned for deleted alarms, but scope changes
# returned to those carried out on behalf of the auth'd tenant, to
# avoid inappropriate cross-tenant visibility of alarm history
auth_project = acl.get_limited_to_project(pecan.request.headers)
conn = pecan.request.alarm_storage_conn
kwargs = _query_to_kwargs(q, conn.get_alarm_changes, ['on_behalf_of',
return [AlarmChange.from_db_model(ac)
for ac in conn.get_alarm_changes(self._id, auth_project,
@wsme_pecan.wsexpose(state_kind_enum, body=state_kind_enum)
def put_state(self, state):
"""Set the state of this alarm.
:param state: an alarm state within the request body.
# note(sileht): body are not validated by wsme
# Workaround for
if state not in state_kind:
raise ClientSideError(_("state invalid"))
now = timeutils.utcnow()
alarm = self._alarm()
alarm.state = state
alarm.state_timestamp = now
alarm = self.conn.update_alarm(alarm)
change = {'state': alarm.state}
self._record_change(change, now, on_behalf_of=alarm.project_id,
return alarm.state
def get_state(self):
"""Get the state of this alarm."""
alarm = self._alarm()
return alarm.state
class AlarmsController(rest.RestController):
"""Manages operations on the alarms collection."""
def _lookup(self, alarm_id, *remainder):
return AlarmController(alarm_id), remainder
def _record_creation(self, conn, data, alarm_id, now):
if not cfg.CONF.alarm.record_history:
type = alarm_models.AlarmChange.CREATION
scrubbed_data = utils.stringify_timestamps(data)
detail = json.dumps(scrubbed_data)
user_id = pecan.request.headers.get('X-User-Id')
project_id = pecan.request.headers.get('X-Project-Id')
payload = dict(event_id=str(uuid.uuid4()),
except ceilometer.NotImplementedError:
# Revert to the pre-json'ed details ...
payload['detail'] = scrubbed_data
_send_notification(type, payload)
@wsme_pecan.wsexpose(Alarm, body=Alarm, status_code=201)
def post(self, data):
"""Create a new alarm.
:param data: an alarm within the request body.
conn = pecan.request.alarm_storage_conn
now = timeutils.utcnow()
data.alarm_id = str(uuid.uuid4())
user_limit, project_limit = acl.get_limited_to(pecan.request.headers)
def _set_ownership(aspect, owner_limitation, header):
attr = '%s_id' % aspect
requested_owner = getattr(data, attr)
explicit_owner = requested_owner != wtypes.Unset
caller = pecan.request.headers.get(header)
if (owner_limitation and explicit_owner
and requested_owner != caller):
raise ProjectNotAuthorized(requested_owner, aspect)
actual_owner = (owner_limitation or
requested_owner if explicit_owner else caller)
setattr(data, attr, actual_owner)
_set_ownership('user', user_limit, 'X-User-Id')
_set_ownership('project', project_limit, 'X-Project-Id')
# Check if there's room for one more alarm
if is_over_quota(conn, data.project_id, data.user_id):
raise OverQuota(data)
data.timestamp = now
data.state_timestamp = now
change = data.as_dict(alarm_models.Alarm)
# make sure alarms are unique by name per project.
alarms = list(conn.get_alarms(,
if alarms:
raise ClientSideError(
_("Alarm with name='%s' exists") %,
alarm_in = alarm_models.Alarm(**change)
except Exception:
LOG.exception(_("Error while posting alarm: %s") % change)
raise ClientSideError(_("Alarm incorrect"))
alarm = conn.create_alarm(alarm_in)
self._record_creation(conn, change, alarm.alarm_id, now)
return Alarm.from_db_model(alarm)
@wsme_pecan.wsexpose([Alarm], [Query])
def get_all(self, q=None):
"""Return all alarms, based on the query provided.
:param q: Filter rules for the alarms to be returned.
q = q or []
# Timestamp is not supported field for Simple Alarm queries
kwargs = _query_to_kwargs(q,
return [Alarm.from_db_model(m)
for m in pecan.request.alarm_storage_conn.get_alarms(**kwargs)]
class TraitDescription(_Base):
"""A description of a trait, with no associated value."""
type = wtypes.text
"the data type, defaults to string"
name = wtypes.text
"the name of the trait"
def sample(cls):
return cls(name='service',
class EventQuery(Query):
"""Query arguments for Event Queries."""
_supported_types = ['integer', 'float', 'string', 'datetime']
type = wsme.wsattr(wtypes.text, default='string')
"the type of the trait filter, defaults to string"
def __repr__(self):
# for logging calls
return '<EventQuery %r %s %r %s>' % (self.field,
class Trait(_Base):
"""A Trait associated with an event."""
name = wtypes.text
"The name of the trait"
value = wtypes.text
"the value of the trait"
type = wtypes.text
"the type of the trait (string, integer, float or datetime)"
def _convert_storage_trait(trait):
"""Helper method to convert a storage model into an API trait instance.
If an API trait instance is passed in, just return it.
if isinstance(trait, Trait):
return trait
value = (six.text_type(trait.value)
if not trait.dtype == event_models.Trait.DATETIME_TYPE
else trait.value.isoformat())
trait_type = event_models.Trait.get_name_by_type(trait.dtype)
return Trait(, type=trait_type, value=value)
def sample(cls):
return cls(name='service',
class Event(_Base):
"""A System event."""
message_id = wtypes.text
"The message ID for the notification"
event_type = wtypes.text
"The type of the event"
_traits = None
def get_traits(self):
return self._traits
def set_traits(self, traits):
self._traits = map(Trait._convert_storage_trait, traits)
traits = wsme.wsproperty(wtypes.ArrayType(Trait),
"Event specific properties"
generated = datetime.datetime
"The time the event occurred"
def sample(cls):
return cls(
'request_id': 'req-4e2d67b8-31a4-48af-bb2f-9df72a353a72',
'service': 'conductor.tem-devstack-01',
'tenant_id': '7f13f2b17917463b9ee21aa92c4b36d6'
def requires_admin(func):
def wrapped(*args, **kwargs):
usr_limit, proj_limit = acl.get_limited_to(pecan.request.headers)
# If User and Project are None, you have full access.
if usr_limit and proj_limit:
# since this decorator get's called out of wsme context
# raising exception results internal error so call abort
# for handling the error
ex = ProjectNotAuthorized(proj_limit)
pecan.core.abort(status_code=ex.code, detail=ex.msg)
return func(*args, **kwargs)
return wrapped
def _event_query_to_event_filter(q):
evt_model_filter = {
'event_type': None,
'message_id': None,
'start_time': None,
'end_time': None
traits_filter = []
for i in q:
if not i.op:
i.op = 'eq'
elif i.op not in operation_kind:
error = _("operator {} is incorrect").format(i.op)
raise ClientSideError(error)
if i.field in evt_model_filter:
evt_model_filter[i.field] = i.value
traits_filter.append({"key": i.field,
i.type: i._get_value_as_type(),
"op": i.op})
return storage.EventFilter(traits_filter=traits_filter, **evt_model_filter)
class TraitsController(rest.RestController):
"""Works on Event Traits."""
@wsme_pecan.wsexpose([Trait], wtypes.text, wtypes.text)
def get_one(self, event_type, trait_name):
"""Return all instances of a trait for an event type.
:param event_type: Event type to filter traits by
:param trait_name: Trait to return values for
LOG.debug(_("Getting traits for %s") % event_type)
return [Trait._convert_storage_trait(t)
for t in pecan.request.event_storage_conn
.get_traits(event_type, trait_name)]
@wsme_pecan.wsexpose([TraitDescription], wtypes.text)
def get_all(self, event_type):
"""Return all trait names for an event type.
:param event_type: Event type to filter traits by
get_trait_name = event_models.Trait.get_name_by_type
return [TraitDescription(name=t['name'],
for t in pecan.request.event_storage_conn
class EventTypesController(rest.RestController):
"""Works on Event Types in the system."""
traits = TraitsController()
def get_one(self, event_type):
def get_all(self):
"""Get all event types."""
return list(pecan.request.event_storage_conn.get_event_types())
class EventsController(rest.RestController):
"""Works on Events."""
@wsme_pecan.wsexpose([Event], [EventQuery])
def get_all(self, q=None):
"""Return all events matching the query filters.
:param q: Filter arguments for which Events to return
q = q or []
event_filter = _event_query_to_event_filter(q)
return [Event(message_id=event.message_id,
for event in
@wsme_pecan.wsexpose(Event, wtypes.text)
def get_one(self, message_id):
"""Return a single event with the given message id.
:param message_id: Message ID of the Event to be returned
event_filter = storage.EventFilter(message_id=message_id)
events = [event for event
in pecan.request.event_storage_conn.get_events(event_filter)]
if not events:
raise EntityNotFound(_("Event"), message_id)
if len(events) > 1:
LOG.error(_("More than one event with "
"id %s returned from storage driver") % message_id)
event = events[0]
return Event(message_id=event.message_id,
class QuerySamplesController(rest.RestController):
"""Provides complex query possibilities for samples."""
@wsme_pecan.wsexpose([Sample], body=ComplexQuery)
def post(self, body):
"""Define query for retrieving Sample data.
:param body: Query rules for the samples to be returned.
sample_name_mapping = {"resource": "resource_id",
"meter": "counter_name",
"type": "counter_type",
"unit": "counter_unit",
"volume": "counter_volume"}
query = ValidatedComplexQuery(body,
conn = pecan.request.storage_conn
return [Sample.from_db_model(s)
for s in conn.query_samples(query.filter_expr,
class QueryAlarmHistoryController(rest.RestController):
"""Provides complex query possibilities for alarm history."""
@wsme_pecan.wsexpose([AlarmChange], body=ComplexQuery)
def post(self, body):
"""Define query for retrieving AlarmChange data.
:param body: Query rules for the alarm history to be returned.
query = ValidatedComplexQuery(body,
conn = pecan.request.alarm_storage_conn
return [AlarmChange.from_db_model(s)
for s in conn.query_alarm_history(query.filter_expr,
class QueryAlarmsController(rest.RestController):
"""Provides complex query possibilities for alarms."""
history = QueryAlarmHistoryController()
@wsme_pecan.wsexpose([Alarm], body=ComplexQuery)
def post(self, body):
"""Define query for retrieving Alarm data.
:param body: Query rules for the alarms to be returned.
query = ValidatedComplexQuery(body,
conn = pecan.request.alarm_storage_conn
return [Alarm.from_db_model(s)
for s in conn.query_alarms(query.filter_expr,
class QueryController(rest.RestController):
samples = QuerySamplesController()
alarms = QueryAlarmsController()