cloudkitty/doc/source/developer/storage.rst

Storage backend (v2)

Warning

This backend is considered unstable and should be used for upstream development only.

In order to implement a storage backend for cloudkitty, you'll have to implement the following abstract class:

cloudkitty.storage.v2.BaseStorage

You'll then need to register an entrypoint corresponding to your storage backend in the cloudkitty.storage.v2.backends section of the setup.cfg file.

Testing

There is a generic test class for v2 storage backends. It allows to run a functional test suite against a new v2 storage backend.

$ tree cloudkitty/tests/storage/v2
cloudkitty/tests/storage/v2
├── base_functional.py
├── __init__.py
└── test_gnocchi_functional.py

In order to use the class, add a file called test_mybackend_functional.py to the cloudkitty/tests/storage/v2 directory. You will then need to write a class inheriting from BaseFunctionalStorageTest. Specify the storage version and the backend name as class attributes

Example:

import testtools

from cloudkitty.tests.storage.v2 import base_functional
from cloudkitty.tests.utils import is_functional_test


@testtools.skipUnless(is_functional_test(), 'Test is not a functional test')
class GnocchiBaseFunctionalStorageTest(
        base_functional.BaseFunctionalStorageTest):

    storage_backend = 'gnocchi'
    storage_version = 2

Two methods need to be implemented:

• wait_for_backend. This method is called once data has been once dataframes have been pushed to the storage backend (in gnocchi's case, it waits for all measures to have been processed). It is a classmethod.
• cleanup_backend: This method is called at the end of the test suite in order to delete all data from the storage backend. It is a classmethod.