how to enable hardware metrics on gnocchi

asked 2017-04-28 15:14:28 -0500

mkhan
21 ●22 ●31 ●35

My current environment is Ocata and gnochi is getting all the metrics except hardware metrics. I have configure snmp and snmpwalk is working fine. Somehow, gnocchi is show any any hardware metrics.

cat /etc/ceilometer/gnocchi.resource.yml

resources: - resource_type: identity archive_policy: low metrics: - 'identity.authenticate.success' - 'identity.authenticate.pending' - 'identity.authenticate.failure' - 'identity.user.created' - 'identity.user.deleted' - 'identity.user.updated' - 'identity.group.created' - 'identity.group.deleted' - 'identity.group.updated' - 'identity.role.created' - 'identity.role.deleted' - 'identity.role.updated' - 'identity.project.created' - 'identity.project.deleted' - 'identity.project.updated' - 'identity.trust.created' - 'identity.trust.deleted' - 'identity.role_assignment.created' - 'identity.role_assignment.deleted'

resource_type: ceph_account metrics:
- 'radosgw.objects'
- 'radosgw.objects.size'
- 'radosgw.objects.containers'
- 'radosgw.api.request'
- 'radosgw.containers.objects'
- 'radosgw.containers.objects.size'
resource_type: instance metrics:
- 'memory'
- 'memory.usage'
- 'memory.resident'
- 'memory.bandwidth.total'
- 'memory.bandwidth.local'
- 'vcpus'
- 'cpu'
- 'cpu.delta'
- 'cpu_util'
- 'cpu_l3_cache'
- 'disk.root.size'
- 'disk.ephemeral.size'
- 'disk.read.requests'
- 'disk.read.requests.rate'
- 'disk.write.requests'
- 'disk.write.requests.rate'
- 'disk.read.bytes'
- 'disk.read.bytes.rate'
- 'disk.write.bytes'
- 'disk.write.bytes.rate'
- 'disk.latency'
- 'disk.iops'
- 'disk.capacity'
- 'disk.allocation'
- 'disk.usage'
- 'compute.instance.booting.time'
- 'perf.cpu.cycles'
- 'perf.instructions'
- 'perf.cache.references'
- 'perf.cache.misses' attributes: host: resource_metadata.(instance_host|host) image_ref: resource_metadata.image_ref display_name: resource_metadata.display_name flavor_id: resource_metadata.(instance_flavor_id|(flavor.id)) server_group: resource_metadata.user_metadata.server_group event_delete: compute.instance.delete.start event_attributes: id: instance_id event_associated_resources: instance_network_interface: '{"=": {"instance_id": "%s"}}' instance_disk: '{"=": {"instance_id": "%s"}}'
resource_type: instance_network_interface metrics:
- 'network.outgoing.packets.rate'
- 'network.incoming.packets.rate'
- 'network.outgoing.packets'
- 'network.incoming.packets'
- 'network.outgoing.packets.drop'
- 'network.incoming.packets.drop'
- 'network.outgoing.packets.error'
- 'network.incoming.packets.error'
- 'network.outgoing.bytes.rate'
- 'network.incoming.bytes.rate'
- 'network.outgoing.bytes'
- 'network.incoming.bytes' attributes: name: resource_metadata.vnic_name instance_id: resource_metadata.instance_id
resource_type: instance_disk metrics:
- 'disk.device.read.requests'
- 'disk.device.read.requests.rate'
- 'disk.device.write.requests'
- 'disk.device.write.requests.rate'
- 'disk.device.read.bytes'
- 'disk.device.read.bytes.rate'
- 'disk.device.write.bytes'
- 'disk.device.write.bytes.rate'
- 'disk.device.latency'
- 'disk.device.iops'
- 'disk.device.capacity'
- 'disk.device.allocation'
- 'disk.device.usage' attributes: name: resource_metadata.disk_name instance_id: resource_metadata.instance_id
resource_type: image metrics:
- 'image.size'
- 'image.download'
- 'image.serve' attributes: name: resource_metadata.name container_format: resource_metadata.container_format disk_format: resource_metadata.disk_format event_delete: image.delete event_attributes: id: resource_id
resource_type: ipmi metrics:
- 'hardware.ipmi.node.power'
- 'hardware.ipmi.node.temperature'
- 'hardware.ipmi.node.inlet_temperature'
- 'hardware.ipmi.node.outlet_temperature'
- 'hardware.ipmi.node.fan'
- 'hardware.ipmi.node.current'
- 'hardware.ipmi.node.voltage'
- 'hardware.ipmi.node.airflow'
- 'hardware.ipmi.node.cups'
- 'hardware.ipmi.node.cpu_util'
- 'hardware.ipmi.node.mem_util'
- 'hardware.ipmi.node.io_util'
resource_type: network metrics:
- 'bandwidth'
- 'network'
- 'network.create'
- 'network.update'
- 'subnet'
- 'subnet.create'
- 'subnet.update'
- 'port'
- 'port.create'
- 'port.update'
- 'router'
- 'router.create'
- 'router.update'
- 'ip.floating'
- 'ip.floating.create'
- 'ip.floating.update' event_delete: floatingip.delete.end event_attributes: id: resource_id
resource_type: stack metrics:
- 'stack.create'
- 'stack.update'
- 'stack.delete'
- 'stack.resume'
- 'stack.suspend'
resource_type: swift_account metrics:
- 'storage.objects.incoming.bytes'
- 'storage.objects.outgoing.bytes'
- 'storage.api.request'
- 'storage.objects.size'
- 'storage.objects'
- 'storage.objects ...

(more)

edit retag flag offensive close merge delete

add a comment

3 answers

Sort by » oldest newest most voted

answered 2017-05-01 04:07:58 -0500

yprokule
601 ●4 ●8

Main question - is ceilometer able to collect this data via snmp ? Have U configured it with correct username/community in ceilometer.conf ?

- name: meter_snmp 
  interval: 60 
  resources: 
    - snmp://10.10.128.46 
    - snmp://10.10.128.47 
    - snmp://10.10.128.48
  meters: 
    - "hardware.cpu"
    - "hardware.memory" 
    - "hardware.disks" 
    - "hardware.network" 
    - "hardware.network" 
    - "hardware.system_stats" 
  sinks: 
    - meter_sink

There are no such meters as: hardware.cpu or hardware.memory, but hardware.cpu.load.1min, hardware.cpu.load.5min, hardware.cpu.load.15min, hardware.cpu.util.

https://docs.openstack.org/admin-guid...

So either explicitly specify meters, or use wildcard, e.g:

 - "hardware.cpu.*"

edit flag offensive delete link

Comments

I have double check /etc/ceilometer/pipeline.yaml on my controller node, it is configured as you recomonded and also gnocchi_resource.yaml has configured as it's mentioned. I am not getting any hardware metrics but snmpwalk is running fine.

mkhan ( 2017-05-01 09:08:06 -0500 )edit

Please enable debug/verbose in ceilometer.conf, restart openstack-ceilometer-* services, and check ceilometer/(central|collector).log

yprokule ( 2017-05-10 04:21:36 -0500 )edit

Did you solve the problem? I've tried setting up however the metrics still do not appear.

pablobrunetti ( 2018-02-02 11:38:46 -0500 )edit

I haven't figured it out yet. I haven't seen any hardware metrics yet.

mkhan ( 2018-03-21 11:07:57 -0500 )edit

I was able to collect the hardware metrics via snmp with Openstack multi node in the Pike version. If you want I'll make another answer on this question.

pablobrunetti ( 2018-03-21 16:07:25 -0500 )edit

see more comments

answered 2019-04-07 09:07:12 -0500

mkhan
21 ●22 ●31 ●35

I have used same polling.yaml and pipeline.yaml file as I have used in Ocata but I am unable collect vm metrics and hardware network metrics.

edit flag offensive delete link

add a comment

answered 2018-03-23 12:36:44 -0500

pablobrunetti

61 ●4 ●6 ●12 https://www.linkedin.c...

updated 2018-03-23 12:38:46 -0500

This configuration is for Openstack Multi node with module Ceilometer and gnocchi. The configurations file are in pastebin.

Controller Configurations

Configure /etc/ceilometer/polling.yaml(https://pastebin.com/raw/wTE5Lfqp) and insert your resources compute snmp.

Configure /etc/ceilometer/pipeline.yaml(https://pastebin.com/raw/a2vmq4KA) and insert your resources compute snmp .

Install the snmp manager and snmp-mibs-downloader:
 - sudo apt install snmp
 - sudo apt install snmp-mibs-downloader
 In /etc/snmp/snmp.conf comment the last line

Compute Configurations

Install the snmpd agent:
 - sudo apt install snmpd

Configure the file /etc/snmp/snmpd.conf(https://pastebin.com/raw/gr0YygqZ) and enable the MIBS used for in your project.

edit flag offensive delete link

Comments

I have followed your recommended changes in pipeline.yaml and polling.yaml files and start showing hardware metrics. I am still missing output of some metrics when I run gnocchi measures show commands.

mkhan ( 2018-04-04 10:45:26 -0500 )edit

PM_NET_OUT(hardware.network.outgoing.bytes.rate) isn't being published PM_NET_IN(hardware.network.incoming.bytes.rate) isn't being published PM_IO_OUT(hardware.system_stats.io.outgoing.blocks.rate) isn't being published PM_IO_IN(hardware.system_stats.io.incoming.blocks.rate) isn't being publis

mkhan ( 2018-04-04 10:45:42 -0500 )edit

memory.used is also not publishing. Thank you so much for your help.

mkhan ( 2018-04-04 10:46:43 -0500 )edit

I suggest you use the MIB Browser software and verify that it is necessary to use new MIBs and verify that the metrics are set up correctly. You perform a GET on your computer to verify that it has been configured correctly.

pablobrunetti ( 2018-04-04 15:57:15 -0500 )edit

these metrics are enable in pipline.yaml and polling.yaml files but not getting vm_memory.used and hardware.network metrics

mkhan ( 2019-04-09 08:32:43 -0500 )edit

add a comment

how to enable hardware metrics on gnocchi

3 answers

Comments

Comments

Your Answer

Get to know Ask OpenStack

Question Tools

Stats

Related questions

Feedback About This Page

OpenStack

Community

Documentation

Branding & Legal

how to enable hardware metrics on gnocchi edit

3 answers

Comments

Comments

Your Answer

Get to know Ask OpenStack

Question Tools

Stats

Related questions

Feedback About This Page

OpenStack

Community

Documentation

Branding & Legal

how to enable hardware metrics on gnocchi