How to enable “memory.usage”, “disk.device.usage” in openstack (ceilometer/nova)
need to get those meter for every 10 minutes (600 secs)
Details :-
Compiled against library: libvirt 1.2.17 Using library: libvirt 1.2.17 Using API: QEMU 1.2.17 Running hypervisor: QEMU 2.6.0
I am new to openstack . Hence could you please post detailed answer/configuration.
I have configured memory usage in pipeline.yaml sources:
[...]
- name: memory_source
interval: 600
meters:
- "memory.usage"
sinks:
- memory_sink
sinks:
[...]
- name: memory_sink
transformers:
- name: "arithmetic"
parameters:
target:
name: "memory_usage"
unit: "%"
type: "gauge"
expr: "100 * $(memory.usage) / ($(memory.usage).resource_metadata.memory_mb)"
However I am unable to get disk.device.usage
Disk.device.usage and memory.usageare built into Ceilometer, see https://docs.openstack.org/ceilometer.... There is nothing you need to do to enable them, except possibly the configuration as documented in the installation guide.
requirement is to get the disk.device.usage in percentage.I have no idea how to do that in pipeline.yaml. But i have done it for memory usage . I have tried same for disk.device.usage and it is not working. disk.device.usage is not coming in the command $ ceilometer meter-list using mongodb
I have no experience crafting pipeline transformers.
Perhaps the meter-list command only prints a limited number of meters. To reduce the output, try a query as documented in https://docs.openstack.org/ceilometer....
Asked: 2018-01-09 00:59:17 -0600
Seen: 761 times
Last updated: Jan 18 '18
See the installation guide at https://docs.openstack.org/ceilometer.... Depending on the way your cloud is deployed, it may be applied as-is, or you may have to adapt it to your situation.
600 sec is the default sampling interval.
Thanks Bernd. memory.usage and disk.device.usage is not configured in pipeline.yaml or nova.conf. I want to know, how to configure those parameters and enable it at openstack level. so that i can query from application for every 600 sec. as of now I can query cpu_util .
Metrics are never configured in nova.conf. Memory usage and disk usage are built into Ceilometer, so that you should not have to configure them in pipeline.yaml either. The installation guide should be sufficient.
Are you using gnocchi or mongodb?
mongodb. I want to get disk.device.usage in percentage.