Skip to content

[RFE] Update metrics collection #879

@falencastro

Description

@falencastro

Since OpenStack Stein (RHOSP 15) some ceilometer metrics were deprecated, like:

  • cpu_util
  • disk.read.requests.rate
  • disk.write.requests.rate
  • disk.read.bytes.rate
  • disk.write.bytes.rate
  • disk.device.read.requests.rate
  • disk.device.write.requests.rate
  • disk.device.read.bytes.rate
  • disk.device.write.bytes.rate

cpu: (1)
disk: (2)(3)

Some of these are being collected in (4).

This prevents cpu and disk metrics from being collected and properly reported, causes the widget "Top CPU Consumers (weekly)" from the default dashboard to come back empty. Also this could be the issue behind #688.

For the disk metrics it should be only a matter of replacing disk.(read|write).bytes with disk.device.(read|write).bytes.
For the cpu_util metric that is gone, it should be replaced with gnocchi rate aggregation on the 'cpu' metric (rate:mean cpu) (5).

(5): "The cpu_util metric was deprecated and removed from Red Hat OpenStack Platform. To obtain the equivalent functionality, use the cumulative cpu metric and an archive policy that includes the rate:mean aggregation method."

1: https://docs.openstack.org/releasenotes/ceilometer/rocky.html
2: https://docs.openstack.org/releasenotes/ceilometer/stein.html
3: https://bugzilla.redhat.com/show_bug.cgi?id=1852332
4: https://github.com/ManageIQ/manageiq-providers-openstack/blob/master/app/models/manageiq/providers/openstack/cloud_manager/metrics_capture.rb
5: https://access.redhat.com/documentation/pt-br/red_hat_openstack_platform/16.1/html-single/auto_scaling_for_instances/index

@miq-bot add-label bug
@miq-bot add-label enhancement

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions