invalid cpu_util on creation

[gordon chung]

when a new instance is created, the cpu_util value generated is consistently way above 0.xx. it seems it's providing utilisation as integer percentage vs decimal.

[gordon chung (chungg)]

marking as medium. i think this only happens on create and never again.

[Toshiaki Takahashi (takahashi-tsc)]

Behavior close to this bug occurs in my environment.
I have 2 questions.

I booted one server and I checked cpu_util with following commands.
The first cpu_util was about 1.54. After that the cpu_util was about 0.02 and it was stable.
Q1. Is this "1.54" close to the image "way above 0.xx"?

$ ceilometer sample-list -m cpu_util
+--------------------------------------+----------+-------+-----------------+------+----------------------------+
| Resource ID | Name | Type | Volume | Unit | Timestamp |
+--------------------------------------+----------+-------+-----------------+------+----------------------------+
| 898dffb4-b9d0-4700-8632-62a11116bd20 | cpu_util | gauge | 0.0251067476142 | % | 2016-06-27T07:31:46.854038 |
| 898dffb4-b9d0-4700-8632-62a11116bd20 | cpu_util | gauge | 1.54619274584 | % | 2016-06-27T07:30:47.109143 |

I also checked cpu time.
Above cpu_util looks correct in terms of following Volume values.
Q2. I think it is high load at start-up server. Is there something wrong in my understand?

$ ceilometer sample-list -m cpu
+--------------------------------------+------+------------+--------------+------+----------------------------+
| Resource ID | Name | Type | Volume | Unit | Timestamp |
+--------------------------------------+------+------------+--------------+------+----------------------------+
| 898dffb4-b9d0-4700-8632-62a11116bd20 | cpu | cumulative | 3900000000.0 | ns | 2016-06-27T07:31:46.854038 |
| 898dffb4-b9d0-4700-8632-62a11116bd20 | cpu | cumulative | 3840000000.0 | ns | 2016-06-27T07:30:47.109143 |
| 898dffb4-b9d0-4700-8632-62a11116bd20 | cpu | cumulative | 130000000.0 | ns | 2016-06-27T07:29:47.123090 |

[Sumant Murke (sumant-murke)]

I tried to boot a vm and then tried checking ceilometer sample-list -m cpu_util but the % is always below 1. Are there any necessary steps to be followed for replicating or any specific configuration ?

[gordon chung (chungg)]

i need to verify this again.

[gordon chung (chungg)]

i'm going to close this. i'm an idiot. cpu_util should be value from 0 to 100