Hi Jonathan,
On your compute host, can you check the configuration in /etc/ceilometer/pipeline.yaml ?
The default configuration contains :
- name: cpu_source
interval: 600
meters:
- "cpu"
sinks:
- cpu_sink
And
- name: cpu_sink
transformers:
- name: "rate_of_change"
parameters:
target:
name: "cpu_util"
unit: "%"
type: "gauge"
scale: "100.0 / (10**9 * (resource_metadata.cpu_number or 1))"
publishers:
- notifier://
If you change the “interval” parameter to the frequency you are after (in seconds) and restart the ceilometer agent I believe that should solve your problem.
Cheers,
Dan.
From: Jonathan Abdiel Gonzalez Valdebenito [mailto:jonathan.abdiel at gmail.com]
Sent: Saturday, 11 February 2017 5:24 AM
To: openstack at lists.openstack.org; openstack-operators at lists.openstack.org
Subject: [Openstack] Ceilometer and cpu_util meter
Hi All!!
We run a public cloud and doing the proper tests to release heat but we hit that the meter cpu_util which it's the one that appears in almost all heat templates sample, it's taking about 10 minutes between samples which as my friend Sergio says unacceptable[1].
Question, does anyone knows if this it's the expected behavior or there's a way to speed up this meter?
Best Regards,
[1] http://lists.openstack.org/pipermail/openstack-operators/2017-February/012610.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20170212/fa0d04f9/attachment.html>