Same metrics are not being listed for two resources of same resource type "INSTANCE"

[Dilip Renkila]

Hi all,

I have a ceilometer problem. I want bill customers based up on some metrics like vcpus, memory etc... My Ceilometer installation is not working as Expected. Lets consider metrics for an Instance

root@ctrl1:~# openstack metric resource show 89f3d7cd-f168-4dec-9e99-f1994eac8588
rating version 2 is not in supported versions: 1
+-----------------------+-------------------------------------------------------------------+
| Field | Value |
+-----------------------+-------------------------------------------------------------------+
| created_by_project_id | 0a49f61d66644fb2a10d664d5b79b1af |
| created_by_user_id | 4a479a2a8c704d3c88e198e86890e3a5 |
| creator | 4a479a2a8c704d3c88e198e86890e3a5:0a49f61d66644fb2a10d664d5b79b1af |
| ended_at | None |
| id | 89f3d7cd-f168-4dec-9e99-f1994eac8588 |
| metrics | cpu: 755350f4-fc69-43ed-a796-6083f2f764f1 |
| | disk.allocation: c39afe66-2858-4a1e-bf63-700ccb685a68 |
| | disk.capacity: b6d5e670-919e-4bb3-af08-006df8b32635 |
| | disk.ephemeral.size: 61ce0b93-6448-46ad-aa87-dd87db5d90fc |
| | disk.read.bytes: cd95bfd7-57af-454f-bc8b-6f29bdf30a7e |
| | disk.read.requests: 3b5034de-1bf6-4bf8-babd-7bd09ee57270 |
| | disk.root.size: d758eb1b-ce01-4218-8d60-7cd69261764b |
| | disk.usage: af4eb4aa-267a-4285-bfdc-7ec25042b562 |
| | disk.write.bytes: b6cefee5-5a14-46cc-aa4f-83ef52b782a5 |
| | disk.write.requests: bf6d4552-35c2-4639-9405-d7c252382a1b |
| | memory.resident: 762f9e75-f66f-4e01-9908-b11447c062c9 |
| | memory.swap.in: e926f3b1-5efc-4be3-8758-13a82b8ca8d4 |
| | memory.swap.out: 95e1c7a6-f9ad-4c64-894a-8239e3ed7820 |
| | memory.usage: c20bc575-1f59-4043-aae1-90ebe9ea3154 |
| | memory: 52438356-775b-4f1c-ba95-86937e8da90a |
| | vcpus: 7f7a960e-3d3d-4b97-9abc-a2583290f43b |
| original_resource_id | 89f3d7cd-f168-4dec-9e99-f1994eac8588 |
| project_id | 3e3d4490d2a1426ea1faa76761476988 |
| revision_end | None |
| revision_start | 2019-01-08T16:05:42.705597+00:00 |
| started_at | 2019-01-08T16:02:25.265544+00:00 |
| type | instance |
| user_id | bf291d2a764240dab3bf4aaa4228ec45 |
+-----------------------+-------------------------------------------------------------------+

and then an another one

root@ctrl1:~# openstack metric resource show 49e8f9b6-b3cb-4bc8-9f64-1df422ad774a
rating version 2 is not in supported versions: 1
+-----------------------+-------------------------------------------------------------------+
| Field | Value |
+-----------------------+-------------------------------------------------------------------+
| created_by_project_id | 0a49f61d66644fb2a10d664d5b79b1af |
| created_by_user_id | 4a479a2a8c704d3c88e198e86890e3a5 |
| creator | 4a479a2a8c704d3c88e198e86890e3a5:0a49f61d66644fb2a10d664d5b79b1af |
| ended_at | None |
| id | 49e8f9b6-b3cb-4bc8-9f64-1df422ad774a |
| metrics | cpu: be495114-fb18-42d3-a7e8-1493d60c69bd |
| | disk.allocation: 536f78f1-0da4-4299-9d52-5f209f9175f9 |
| | disk.capacity: b54fa4cc-152e-4d19-832e-f3e0b82238be |
| | disk.read.bytes: d4781cc0-1913-4c57-8f41-ec1a39049e1b |
| | disk.read.requests: 5ddb7caf-1ffb-416a-90c4-b8fd7f59e236 |
| | disk.usage: 5ddf9715-1d0a-4f2c-992b-ed2768a2aada |
| | disk.write.bytes: 06156dbe-c684-44a5-ac91-459af4c1f2b4 |
| | disk.write.requests: 996ab593-eaa8-4f7c-8339-255e9ecf2809 |
| | memory.resident: d316c26d-e832-4e8a-9b65-ca378b28b1f8 |
| original_resource_id | 49e8f9b6-b3cb-4bc8-9f64-1df422ad774a |
| project_id | a1dd449c9ce64345af2a7fb05c4aa21f |
| revision_end | None |
| revision_start | 2019-01-08T16:02:54.141879+00:00 |
| started_at | 2019-01-08T16:02:54.141856+00:00 |
| type | instance |
| user_id | 1f7840b022f84efeaffbd9df5361923e |
+-----------------------+-------------------------------------------------------------------+

The former don't have vcpus, memory etcc. The fun part is both of these instances run on same compute host. Is this behavior is normal ? What else could be wrong ?

[Prankul Mahajan (prankul)]

It is normal to not get all the metrics corresponding to that resource instantly.
I believe it takes time for the resource to collect the valid metrics corresponding to it.
I suggest you to run the "resource show" command again and check.