Files

Luka Peschke d070f6a68c Rewrite of the client

The client has been completely rewritten in order to use cliff. The code
should be easier to maintain: authentication is now entirely handled by
keystoneauth, CloudKitty's client and CK's OSC plugin use the exact same
classes (no code duplication).

New features for users:

  * Client-side CSV report generation: It is possible for users to generate
    CSV reports with the new client. There is a default format, but reports
    may also be configured through a yaml config file. (see documentation)

  * The documentation has been improved. (A few examples on how to use the
    python library + complete API bindings and CLI reference).

  * It is now possible to use the client without Keystone authentication (this
    requires that CK's API is configured to use the noauth auth strategy).

  * Various features are brought by cliff: completion, command output formatting
    (table, shell, yaml, json...).

New features for developpers:

  * Python 2.7/3.5 compatible 'python-cloudkittyclient' module.

  * Integration tests (for 'openstack rating' and 'cloudkitty') have been
    added. These allow to create gate jobs running against a CK devstack

  * Tests are now ran with stestr instead of testr, which allows a better
    control over execution.

  * The dependency list has been reduced and upper constraints have been set.

Change-Id: I7c6afa46138d499b37b8be3d049b23ab5302a928
Task: 6589
Story: 2001614

2018-06-15 12:08:21 +02:00

3.8 KiB

Raw Blame History

Usage

Python library

You can use cloudkittyclient with or without keystone authentication. In order to use it without keystone authentication, cloudkittyclient provides the CloudKittyNoAuthPlugin keystoneauth plugin:

>>> from cloudkittyclient import client as ck_client
>>> from cloudkittyclient import auth as ck_auth

>>> auth = ck_auth.CloudKittyNoAuthPlugin(endpoint='http://127.0.0.1:8889')
>>> client = ck_client.Client('1', auth=auth)
>>> client.report.get_summary()
{u'summary': [{u'begin': u'2018-03-01T00:00:00',
               u'end': u'2018-04-01T00:00:00',
               u'rate': u'1672.71269',
               u'res_type': u'ALL',
               u'tenant_id': u'bea6a24f77e946b0a92dca7c78b7870b'}]}

Else, use it the same way as any other OpenStack client:

>>> import os

>>> from keystoneauth1 import session
>>> from keystoneauth1.identity import v3

>>> from cloudkittyclient import client as ck_client

>>> auth = v3.Password(
       auth_url=os.environ.get('OS_AUTH_URL'),
       project_domain_id=os.environ.get('OS_PROJECT_DOMAIN_ID'),
       user_domain_id=os.environ.get('OS_USER_DOMAIN_ID'),
       username=os.environ.get('OS_USERNAME'),
       project_name=os.environ.get('OS_PROJECT_NAME'),
       password=os.environ.get('OS_PASSWORD'))

>>> ck_session = session.Session(auth=auth)

>>> c = ck_client.Client('1', session=ck_session)

>>> c.report.get_summary()
{u'summary': [{u'begin': u'2018-03-01T00:00:00',
                u'end': u'2018-04-01T00:00:00',
                u'rate': u'1672.71269',
                u'res_type': u'ALL',
                u'tenant_id': u'bea6a24f77e946b0a92dca7c78b7870b'}]}

When using the cloudkitty CLI client with keystone authentication, the auth plugin to use should automagically be detected. If not, you can specify the auth plugin to use with --os-auth-type/--os-auth-plugin:

$ cloudkitty --debug --os-auth-type cloudkitty-noauth summary get
+------------+---------------+------------+---------------------+---------------------+
| Project ID | Resource Type | Rate       | Begin Time          | End Time            |
+------------+---------------+------------+---------------------+---------------------+
| ALL        | ALL           | 1676.95499 | 2018-03-01T00:00:00 | 2018-04-01T00:00:00 |
+------------+---------------+------------+---------------------+---------------------+

CSV report generation

An output formatter (DataframeToCsvFormatter) has been created in order to allow CSV report generation through the client. It can be used with the -f df-to-csv option.

$ cloudkitty dataframes get -b 2018-03-22T12:00:00 -f df-to-csv
Begin,End,Metric Type,Qty,Cost,Project ID,Resource ID,User ID
2018-03-01T12:00:00,2018-03-01T13:00:00,compute,1,2.0,53c3fe396a1a4ab0914b9aa997a5ff88,382d23c3-7b77-4e32-8d65-b3baf86ed7bb,38c1949c2e624f729b30e034ac787640
[...]

Warning

The df-to-csv formatter should NEVER be used together with the -c/--column option and should only be used for the dataframes get command.

The example above shows how to get a CSV report with the standard columns. If you want other columns, it is possible to customize the formatter through a configuration file:

../../etc/cloudkitty/csv_config.yml

Example with this config file:

$ cloudkitty dataframes get -f df-to-csv --format-config-file /etc/cloudkitty/csv_config.yml > report.csv
$ head -n 2 report.csv
Begin,End,User ID,Resource ID,Qty,Cost
2018-03-01T12:00:00,2018-03-01T13:00:00,38c1949c2e624f729b30e034ac787640,382d23c3-7b77-4e32-8d65-b3baf86ed7bb,1,2.0

An other config file is provided: legacy_csv_config.yml. This file is compatible with the format of cloudkitty-writer's CSV reports.

3.8 KiB Raw Blame History

Usage

Python library

CSV report generation

3.8 KiB

Raw Blame History