OpenStack ManualsOpenStack End User Guide  - current

  Atom feed of this document
  
 

 Measure cloud resources

Telemetry measures cloud resources in OpenStack.

It collects information about how much, who, what, and when with regards to billing. Currently, metering is available through only the ceilometer command-line client.

To model data, Telemetry uses these abstractions:

Meter

Measures a specific aspect of resource usage, such as the existence of a running instance, or ongoing performance, such as the CPU utilization for an instance. Meters exist for each type of resource. For example, a separate cpu_util meter exists for each instance. The life cycle of a meter is decoupled from the existence of its related resource. The meter persists after the resource goes away.

A meter has the following attributes:

  • String name.

  • A unit of measurement.

  • A type. Indicates whether values increase monotonically (cumulative), are interpreted as a change from the previous value (delta), or are standalone and relate only to the current duration (gauge).

Sample

An individual data point that is associated with a specific meter. Has the same attributes as the meter, with the addition of timestamp and value attributes. The value attribute is also known as the sample volume.

Statistic

A set of data point aggregates over a time duration. (In contrast, a sample represents a single data point.) The Telemetry service employs these aggregation functions:

  • count. The number of samples in each period.

  • max. The maximum number of sample volumes in each period.

  • min. The minimum number of sample volumes in each period.

  • avg. The average of sample volumes over each period.

  • sum. The sum of sample volumes over each period.

Alarm

A set of rules that define a monitor and a current state, with edge-triggered actions associated with target states. Provides user-oriented Monitoring-as-a-Service and a general purpose utility for OpenStack. Orchestration auto scaling is a typical use-case. Alarms follow a tristate model of ok, alarm, and insufficient data. For conventional threshold-oriented alarms, a static threshold value and comparison operator govern state transitions. The comparison operator compares a selected meter statistic against an evaluation window of configurable length into the recent past.

This example uses the heat client to create an auto-scaling stack and the ceilometer client to measure resources.

  1. Create an auto-scaling stack:

    $ heat stack-create -f cfn/F17/AutoScalingCeilometer.yaml -P "KeyName=heat_key"

  2. List the heat resources that were created:

    $ heat resource-list
+--------------------------+-----------------------------------------+-----------------+----------------------+
| resource_name            | resource_type                           |resource_status | updated_time         |
+--------------------------+-----------------------------------------+-----------------+----------------------+
| CfnUser                  | AWS::IAM::User                          |CREATE_COMPLETE | 2013-10-02T05:53:41Z |
| WebServerKeys            | AWS::IAM::AccessKey                     |CREATE_COMPLETE | 2013-10-02T05:53:42Z |
| LaunchConfig             | AWS::AutoScaling::LaunchConfiguration   |CREATE_COMPLETE | 2013-10-02T05:53:43Z |
| ElasticLoadBalancer      | AWS::ElasticLoadBalancing::LoadBalancer |UPDATE_COMPLETE | 2013-10-02T05:55:58Z |
| WebServerGroup           | AWS::AutoScaling::AutoScalingGroup      |CREATE_COMPLETE | 2013-10-02T05:55:58Z |
| WebServerScaleDownPolicy | AWS::AutoScaling::ScalingPolicy         |CREATE_COMPLETE | 2013-10-02T05:56:00Z |
| WebServerScaleUpPolicy   | AWS::AutoScaling::ScalingPolicy         |CREATE_COMPLETE | 2013-10-02T05:56:00Z |
| CPUAlarmHigh             | OS::Ceilometer::Alarm                   |CREATE_COMPLETE | 2013-10-02T05:56:02Z |
| CPUAlarmLow              | OS::Ceilometer::Alarm                   |CREATE_COMPLETE | 2013-10-02T05:56:02Z |
+--------------------------+-----------------------------------------+-----------------+----------------------+

  3. List the alarms that are set:

    $ ceilometer alarm-list
+--------------------------------------+------------------------------+-------------------+---------+------------+-------------------------------+
| Alarm ID                             | Name                         | State             | Enabled | Continuous | Alarm condition               |
+--------------------------------------+------------------------------+-------------------+---------+------------+-------------------------------+
| 4f896b40-0859-460b-9c6a-b0d329814496 | as-CPUAlarmLow-i6qqgkf2fubs  | insufficient data | True    | False      | cpu_util < 15.0 during 1x 60s |
| 75d8ecf7-afc5-4bdc-95ff-19ed9ba22920 | as-CPUAlarmHigh-sf4muyfruy5m | insufficient data | True    | False      | cpu_util > 50.0 during 1x 60s |
+--------------------------------------+------------------------------+-------------------+---------+------------+-----------------------------+

  4. List the meters that are set:

    $ ceilometer meter-list
+--------------------------+------------+----------+-----------------------------------------------------------------------+------------------------------------+
| Name                     | Type       | Unit     | Resource ID                          | User ID                          | Project ID                       |
+--------------------------+------------+----------+-----------------------------------------------------------------------+------------------------------------+
| cpu                      | cumulative | ns       | 3965b41b-81b0-4386-bea5-6ec37c8841c1 | d1a2996d3b1f4e0e8645ba9650308011 | bf03bf32e3884d489004ac995ff7a61c |
| cpu                      | cumulative | ns       | 62520a83-73c7-4084-be54-275fe770ef2c | d1a2996d3b1f4e0e8645ba9650308011 | bf03bf32e3884d489004ac995ff7a61c |
| cpu_util                 | gauge      | %        | 3965b41b-81b0-4386-bea5-6ec37c8841c1 | d1a2996d3b1f4e0e8645ba9650308011 | bf03bf32e3884d489004ac995ff7a61c |
+--------------------------+------------+----------+-----------------------------------------------------------------------+------------------------------------+

  5. List samples:

    $ ceilometer sample-list -m cpu_util
+--------------------------------------+----------+-------+---------------+------+---------------------+
| Resource ID                          | Name     | Type  | Volume        | Unit | Timestamp           |
+--------------------------------------+----------+-------+---------------+------+---------------------+
| 3965b41b-81b0-4386-bea5-6ec37c8841c1 | cpu_util | gauge | 3.98333333333 | %    | 2013-10-02T10:50:12 |
+--------------------------------------+----------+-------+---------------+------+---------------------+

  6. View statistics:

    $ ceilometer statistics -m cpu_util
+--------+---------------------+---------------------+-------+---------------+---------------+---------------+---------------+----------+---------------------+---------------------+
| Period | Period Start        | Period End          | Count | Min           | Max           | Sum           | Avg           | Duration | Duration Start      | Duration End        |
+--------+---------------------+---------------------+-------+---------------+---------------+---------------+---------------+----------+---------------------+---------------------+
| 0      | 2013-10-02T10:50:12 | 2013-10-02T10:50:12 | 1     | 3.98333333333 | 3.98333333333 | 3.98333333333 | 3.98333333333 | 0.0      | 2013-10-02T10:50:12 | 2013-10-02T10:50:12 |
+--------+---------------------+---------------------+-------+---------------+---------------+---------------+---------------+----------+---------------------+---------------------+
Questions? Discuss on ask.openstack.org
Found an error? Report a bug against this page
Legal notices
Prev Up Next
 Home 
loading table of contents...