yahoo-eng-team team mailing list archive

[Rodrigo Barbieri <1915055@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

environment: devstack-master stacked Jan 28th 2021

The "openstack usage show" commands provide metrics related to run time of the instance, as seen below:
+---------------+----------+
| Field         | Value    |
+---------------+----------+
| CPU Hours     | 260.68   |
| Disk GB-Hours | 260.68   |
| RAM MB-Hours  | 66733.63 |
| Servers       | 3        |
+---------------+----------+

The logic in [0] determines how those values are calculated. They are
based on the launched_at and terminated_at fields.

Some operations, such as resize and unshelve, reset the launched_at
field. Therefore, for a given instance, the run time information is
wiped, as if it had never run before.

Steps to reproduce:
1. Create an instance.
2. Wait a few minutes, start monitoring usage with "watch openstack usage show --project admin" on a separate tab.
3. Either shelve and unshelve an instance, or resize the instance and revert the resize.
4. Notice how the "openstack usage show" statistics suddenly drops to a lower value and then continues to increase.

Expected result:
Statistics would not drop, should continue measuring.

Some possible solutions:
1. Stop resetting the launched_at field
2. Change the field used for calculation at [0] to something else (maybe created_at?)

[0]
https://github.com/openstack/nova/blob/6c0ceda3659405149b7c0b5c283275ef0a896269/nova/api/openstack/compute/simple_tenant_usage.py#L74

** Affects: nova
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1915055

Title:
  launched_at's reset when resizing/reverting and unshelving impacts
  "openstack usage show"

Status in OpenStack Compute (nova):
  New

Bug description:
  environment: devstack-master stacked Jan 28th 2021

  The "openstack usage show" commands provide metrics related to run time of the instance, as seen below:
  +---------------+----------+
  | Field         | Value    |
  +---------------+----------+
  | CPU Hours     | 260.68   |
  | Disk GB-Hours | 260.68   |
  | RAM MB-Hours  | 66733.63 |
  | Servers       | 3        |
  +---------------+----------+

  The logic in [0] determines how those values are calculated. They are
  based on the launched_at and terminated_at fields.

  Some operations, such as resize and unshelve, reset the launched_at
  field. Therefore, for a given instance, the run time information is
  wiped, as if it had never run before.

  Steps to reproduce:
  1. Create an instance.
  2. Wait a few minutes, start monitoring usage with "watch openstack usage show --project admin" on a separate tab.
  3. Either shelve and unshelve an instance, or resize the instance and revert the resize.
  4. Notice how the "openstack usage show" statistics suddenly drops to a lower value and then continues to increase.

  Expected result:
  Statistics would not drop, should continue measuring.

  Some possible solutions:
  1. Stop resetting the launched_at field
  2. Change the field used for calculation at [0] to something else (maybe created_at?)

  [0]
  https://github.com/openstack/nova/blob/6c0ceda3659405149b7c0b5c283275ef0a896269/nova/api/openstack/compute/simple_tenant_usage.py#L74

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1915055/+subscriptions