-
Story
-
Resolution: Done
-
Critical
-
None
-
None
-
2
-
False
-
None
-
False
-
Testable
-
No
-
-
-
-
-
-
-
No
-
No
-
Pending
-
None
-
-
-
ML Ops Sprint 1.20
As part of the testing for enabling RHODS in AWS marketplace, we discovered that the recording metric for vCPU-hours is off by a factor of 2. We believe it's because the metric is counting physical cores instead of vCPUs and 1 core = 2 vCPUs. We are using the same metric as OSD, but we believe the OSD metric is off. For this story, we need to implement a new recording metric that accounts for 1 core=2vCPUs. This should be a straightforward derivative of the existing recording metric.