ceilometer statistics timeout
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Won't Fix
|
Medium
|
MOS Ceilometer | ||
7.0.x |
Won't Fix
|
Medium
|
MOS Maintenance | ||
8.0.x |
Won't Fix
|
Medium
|
MOS Maintenance | ||
Mitaka |
Won't Fix
|
Medium
|
MOS Ceilometer |
Bug Description
Hi
there is a strange ceilometer-api behavior. While I can query ceilometer-api for "meter-
[user1@ubuntu ~]$ ceilometer statistics -q "project=
b'Gateway Timeout (HTTP 504)'
Environment:
MOS 8.0 + Mantainance Update 1
ubuntu 14.04 LTS
I have applied https:/
Some ceilometer-api tests (with the time it takes to be completed):
[user1@ubuntu ~]$ time openstack usage show
Usage from 2016-06-02 to 2016-07-01 on project 6294315f9df1491
+------
| Field | Value |
+------
| CPU Hours | 164.21 |
| Disk GB-Hours | 16420.51 |
| RAM MB-Hours | 168146.07 |
| Servers | 1 |
+------
real 0m2.204s
user 0m1.490s
sys 0m0.113s
[user1@ubuntu ~]$ time ceilometer meter-list
+------
| Name | Type | Unit | Resource ID | User ID | Project ID |
+------
| cpu | cumulative | ns | 5c305833-
| cpu.delta | delta | ns | 5c305833-
.............
| vcpus | gauge | vcpu | 5c305833-
+------
real 0m2.088s
user 0m0.897s
sys 0m0.057s
[user1@ubuntu ~]$ time ceilometer sample-list -q "timestamp>
+------
| ID | Resource ID | Name | Type | Volume | Unit | Timestamp |
+------
| e81ed112-
......
| 9b50bb52-
+------
real 0m2.373s
user 0m0.977s
sys 0m0.073s
Changed in fuel: | |
milestone: | 10.0 → 9.1 |
Changed in fuel: | |
milestone: | 9.1 → 8.0-updates |
Changed in fuel: | |
status: | Incomplete → Confirmed |
tags: |
added: area-ceilometer removed: area-mos ceilometer |
Changed in fuel: | |
milestone: | 8.0-updates → 9.2 |
Changed in fuel: | |
status: | Fix Committed → Confirmed |
Changed in fuel: | |
status: | Confirmed → Won't Fix |
Hi
new update: after rebooting all my three controllers, now I can get the expected result, but it takes a loooong time, too much time actually:
time ceilometer statistics -q "project= 6294315f9df1491 fa9345701d2642c 96;timestamp> 2016-06- 01T00:00" -m vcpus --period 60 --+---- ------- ------- ---+--- ------- ------- ----+-- ---+--- --+---- -+----- +------ -+----- -----+- ------- ------- ------- ------+ ------- ------- ------- ------- + --+---- ------- ------- ---+--- ------- ------- ----+-- ---+--- --+---- -+----- +------ -+----- -----+- ------- ------- ------- ------+ ------- ------- ------- ------- + 30T11:00: 11.167000 | 2016-06- 30T11:00: 11.167000 | 30T12:00: 05.296000 | 2016-06- 30T12:00: 05.296000 | 30T13:00: 56.159000 | 2016-06- 30T13:00: 56.159000 | --+---- ------- ------- ---+--- ------- ------- ----+-- ---+--- --+---- -+----- +------ -+----- -----+- ------- ------- ------- ------+ ------- ------- ------- ------- +
+------
| Period | Period Start | Period End | Max | Min | Avg | Sum | Count | Duration | Duration Start | Duration End |
+------
| 60 | 2016-06-30T11:00:00 | 2016-06-30T11:01:00 | 1.0 | 1.0 | 1.0 | 1.0 | 1 | 0.0 | 2016-06-
| 60 | 2016-06-30T12:00:00 | 2016-06-30T12:01:00 | 1.0 | 1.0 | 1.0 | 1.0 | 1 | 0.0 | 2016-06-
| 60 | 2016-06-30T13:00:00 | 2016-06-30T13:01:00 | 1.0 | 1.0 | 1.0 | 1.0 | 1 | 0.0 | 2016-06-
+------
real 0m22.567s
user 0m0.883s
sys 0m0.090s
22sec with just some instances running (less than 20vm running) in the cloud.
So, probably this is not a bug, but I think that the whole ceilometer/mongodb configuration has some performance problem because all my environment run with new hardware (controllers with > 24GB RAM and SSD samsun pro)