Hello, sorry for the delays on this, it had been waiting on the
customer.

I've updated the template to reflect the user impact, which is that
several of Watcher's audits are unable to provide meaningful actions
because the deprecated metric they're polling doesn't exist and so
Watcher sees 'None' rather than a nominal value which it can use to
generate action plans.

To test the proposed package I deployed an openstack antelope cloud with
gnocchi and watcher overlays.

I then created an Watcher audit based on the workload_balance strategy. I chose 
this strategy because it depends on a metric internal to watcher known as 
instance_cpu_usage, which is intended to be the cpu utilization as a percentage 
[1]
openstack optimize audit create -t CONTINUOUS -i 60 -g workload_balancing -s 
workload_balance --auto-trigger

Then I launched a VM in the cloud so that ceilometer, gnocchi, and watcher 
would retrieve some metrics. Upon running
openstack metric list shows that there is no cpu_util (since it was removed in 
Stein)

The watcher audits show that the instance_cpu_usage is None, and
stressing the VM (artificial workload with stress command) produces no
variance in the polled results

After upgrading to the patched watcher and stressing the VM, the Watcher
audits reflect the detected workload

I repeated the same testing for bobcat

[1]
https://docs.openstack.org/watcher/2023.1/strategies/workload_balance.html

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2088620

Title:
  [SRU] Deprecated usage of cpu_util

To manage notifications about this bug go to:
https://bugs.launchpad.net/cloud-archive/+bug/2088620/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to