Virt inspector directly layered over hypervisor API
Addresses BP nova-independent-virt Add the concept of a pluggable virt inspector that allows polling of the hypervisor layer without relying on any nova internals. Note that the test-requires dependency on the nova master tarball remains in place for the moment, as it's required by the nova notifier tests. Also we must leave the dependency on the libvirt-python RPM or python-libvirt Debian package to the distros which end up packaging ceilometer. Change-Id: I20700320dd7e3196507173c780ab598b479e4021
This commit is contained in:
Eoghan Glynn
269
ceilometer/compute/pollsters.py
Normal file
269
ceilometer/compute/pollsters.py
Normal file
@@ -0,0 +1,269 @@
|
||||
# -*- encoding: utf-8 -*-
|
||||
#
|
||||
# Copyright © 2012 eNovance <licensing@enovance.com>
|
||||
# Copyright © 2012 Red Hat, Inc
|
||||
#
|
||||
# Author: Julien Danjou <julien@danjou.info>
|
||||
# Author: Eoghan Glynn <eglynn@redhat.com>
|
||||
#
|
||||
# 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.
|
||||
|
||||
import copy
|
||||
import datetime
|
||||
|
||||
from stevedore import driver
|
||||
|
||||
from ceilometer import counter
|
||||
from ceilometer.compute import plugin
|
||||
from ceilometer.compute import instance as compute_instance
|
||||
from ceilometer.compute.virt import inspector as virt_inspector
|
||||
from ceilometer.openstack.common import cfg
|
||||
from ceilometer.openstack.common import importutils
|
||||
from ceilometer.openstack.common import log
|
||||
from ceilometer.openstack.common import timeutils
|
||||
|
||||
OPTS = [
|
||||
cfg.StrOpt('hypervisor_inspector',
|
||||
help='Inspector to use for inspecting the hypervisor layer',
|
||||
default='libvirt')
|
||||
]
|
||||
|
||||
CONF = cfg.CONF
|
||||
CONF.register_opts(OPTS)
|
||||
|
||||
LOG = log.getLogger(__name__)
|
||||
|
||||
|
||||
def _instance_name(instance):
|
||||
"""Shortcut to get instance name"""
|
||||
return getattr(instance, 'OS-EXT-SRV-ATTR:instance_name', None)
|
||||
|
||||
|
||||
def get_hypervisor_inspector():
|
||||
try:
|
||||
namespace = 'ceilometer.compute.virt'
|
||||
mgr = driver.DriverManager(namespace,
|
||||
CONF.hypervisor_inspector,
|
||||
invoke_on_load=True)
|
||||
return mgr.driver
|
||||
except ImportError as e:
|
||||
LOG.error("Unable to load the hypervisor inspector: %s" % (e))
|
||||
return virt_inspector.Inspector()
|
||||
|
||||
|
||||
def make_counter_from_instance(instance, name, type, volume):
|
||||
return counter.Counter(
|
||||
name=name,
|
||||
type=type,
|
||||
volume=volume,
|
||||
user_id=instance.user_id,
|
||||
project_id=instance.tenant_id,
|
||||
resource_id=instance.id,
|
||||
timestamp=timeutils.isotime(),
|
||||
resource_metadata=compute_instance.get_metadata_from_object(instance),
|
||||
)
|
||||
|
||||
|
||||
class HypervisorPollster(plugin.ComputePollster):
|
||||
|
||||
inspector = None
|
||||
|
||||
def __init__(self):
|
||||
if not HypervisorPollster.inspector:
|
||||
HypervisorPollster.inspector = get_hypervisor_inspector()
|
||||
|
||||
|
||||
class InstancePollster(HypervisorPollster):
|
||||
|
||||
def get_counters(self, manager, instance):
|
||||
yield make_counter_from_instance(instance,
|
||||
name='instance',
|
||||
type=counter.TYPE_GAUGE,
|
||||
volume=1,
|
||||
)
|
||||
yield make_counter_from_instance(instance,
|
||||
name='instance:%s' %
|
||||
instance.flavor['name'],
|
||||
type=counter.TYPE_GAUGE,
|
||||
volume=1,
|
||||
)
|
||||
|
||||
|
||||
class DiskIOPollster(HypervisorPollster):
|
||||
|
||||
LOG = log.getLogger(__name__ + '.diskio')
|
||||
|
||||
DISKIO_USAGE_MESSAGE = ' '.join(["DISKIO USAGE:",
|
||||
"%s %s:",
|
||||
"read-requests=%d",
|
||||
"read-bytes=%d",
|
||||
"write-requests=%d",
|
||||
"write-bytes=%d",
|
||||
"errors=%d",
|
||||
])
|
||||
|
||||
def get_counters(self, manager, instance):
|
||||
instance_name = _instance_name(instance)
|
||||
try:
|
||||
r_bytes = 0
|
||||
r_requests = 0
|
||||
w_bytes = 0
|
||||
w_requests = 0
|
||||
for disk, info in self.inspector.inspect_disks(instance_name):
|
||||
self.LOG.info(self.DISKIO_USAGE_MESSAGE,
|
||||
instance, disk.device, info.read_requests,
|
||||
info.read_bytes, info.write_requests,
|
||||
info.write_bytes, info.errors)
|
||||
r_bytes += info.read_bytes
|
||||
r_requests += info.read_requests
|
||||
w_bytes += info.write_bytes
|
||||
w_requests += info.write_requests
|
||||
yield make_counter_from_instance(instance,
|
||||
name='disk.read.requests',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=r_requests,
|
||||
)
|
||||
yield make_counter_from_instance(instance,
|
||||
name='disk.read.bytes',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=r_bytes,
|
||||
)
|
||||
yield make_counter_from_instance(instance,
|
||||
name='disk.write.requests',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=w_requests,
|
||||
)
|
||||
yield make_counter_from_instance(instance,
|
||||
name='disk.write.bytes',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=w_bytes,
|
||||
)
|
||||
except Exception as err:
|
||||
self.LOG.warning('Ignoring instance %s: %s',
|
||||
instance_name, err)
|
||||
self.LOG.exception(err)
|
||||
|
||||
|
||||
class CPUPollster(HypervisorPollster):
|
||||
|
||||
LOG = log.getLogger(__name__ + '.cpu')
|
||||
|
||||
utilization_map = {}
|
||||
|
||||
def get_cpu_util(self, instance, cpu_info):
|
||||
prev_times = self.utilization_map.get(instance.id)
|
||||
self.utilization_map[instance.id] = (cpu_info.time,
|
||||
datetime.datetime.now())
|
||||
cpu_util = 0.0
|
||||
if prev_times:
|
||||
prev_cpu = prev_times[0]
|
||||
prev_timestamp = prev_times[1]
|
||||
delta = self.utilization_map[instance.id][1] - prev_timestamp
|
||||
elapsed = (delta.seconds * (10 ** 6) + delta.microseconds) * 1000
|
||||
cores_fraction = 1.0 / cpu_info.number
|
||||
# account for cpu_time being reset when the instance is restarted
|
||||
time_used = (cpu_info.time - prev_cpu
|
||||
if prev_cpu <= cpu_info.time else cpu_info.time)
|
||||
cpu_util = 100 * cores_fraction * time_used / elapsed
|
||||
return cpu_util
|
||||
|
||||
def get_counters(self, manager, instance):
|
||||
self.LOG.info('checking instance %s', instance.id)
|
||||
instance_name = _instance_name(instance)
|
||||
try:
|
||||
cpu_info = self.inspector.inspect_cpus(instance_name)
|
||||
self.LOG.info("CPUTIME USAGE: %s %d",
|
||||
instance.__dict__, cpu_info.time)
|
||||
cpu_util = self.get_cpu_util(instance, cpu_info)
|
||||
self.LOG.info("CPU UTILIZATION %%: %s %0.2f",
|
||||
instance.__dict__, cpu_util)
|
||||
# FIXME(eglynn): once we have a way of configuring which measures
|
||||
# are published to each sink, we should by default
|
||||
# disable publishing this derived measure to the
|
||||
# metering store, only publishing to those sinks
|
||||
# that specifically need it
|
||||
yield make_counter_from_instance(instance,
|
||||
name='cpu_util',
|
||||
type=counter.TYPE_GAUGE,
|
||||
volume=cpu_util,
|
||||
)
|
||||
yield make_counter_from_instance(instance,
|
||||
name='cpu',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=cpu_info.time,
|
||||
)
|
||||
except Exception as err:
|
||||
self.LOG.error('could not get CPU time for %s: %s',
|
||||
instance.id, err)
|
||||
self.LOG.exception(err)
|
||||
|
||||
|
||||
class NetPollster(HypervisorPollster):
|
||||
|
||||
LOG = log.getLogger(__name__ + '.net')
|
||||
|
||||
NET_USAGE_MESSAGE = ' '.join(["NETWORK USAGE:", "%s %s:", "read-bytes=%d",
|
||||
"write-bytes=%d"])
|
||||
|
||||
@staticmethod
|
||||
def make_vnic_counter(instance, name, type, volume, vnic_data):
|
||||
metadata = copy.copy(vnic_data)
|
||||
resource_metadata = dict(zip(metadata._fields, metadata))
|
||||
resource_metadata['instance_id'] = instance.id
|
||||
|
||||
return counter.Counter(
|
||||
name=name,
|
||||
type=type,
|
||||
volume=volume,
|
||||
user_id=instance.user_id,
|
||||
project_id=instance.tenant_id,
|
||||
resource_id=vnic_data.fref,
|
||||
timestamp=timeutils.isotime(),
|
||||
resource_metadata=resource_metadata
|
||||
)
|
||||
|
||||
def get_counters(self, manager, instance):
|
||||
instance_name = _instance_name(instance)
|
||||
self.LOG.info('checking instance %s', instance.id)
|
||||
try:
|
||||
for vnic, info in self.inspector.inspect_vnics(instance_name):
|
||||
self.LOG.info(self.NET_USAGE_MESSAGE, instance_name,
|
||||
vnic.name, info.rx_bytes, info.tx_bytes)
|
||||
yield self.make_vnic_counter(instance,
|
||||
name='network.incoming.bytes',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=info.rx_bytes,
|
||||
vnic_data=vnic,
|
||||
)
|
||||
yield self.make_vnic_counter(instance,
|
||||
name='network.outgoing.bytes',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=info.tx_bytes,
|
||||
vnic_data=vnic,
|
||||
)
|
||||
yield self.make_vnic_counter(instance,
|
||||
name='network.incoming.packets',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=info.rx_packets,
|
||||
vnic_data=vnic,
|
||||
)
|
||||
yield self.make_vnic_counter(instance,
|
||||
name='network.outgoing.packets',
|
||||
type=counter.TYPE_CUMULATIVE,
|
||||
volume=info.tx_packets,
|
||||
vnic_data=vnic,
|
||||
)
|
||||
except Exception as err:
|
||||
self.LOG.warning('Ignoring instance %s: %s',
|
||||
instance_name, err)
|
||||
self.LOG.exception(err)
|
||||
Reference in New Issue
Block a user