Skip to content

Latest commit

 

History

History
79 lines (61 loc) · 2.41 KB

USAGE.md

File metadata and controls

79 lines (61 loc) · 2.41 KB
Raw

Usage

First you need a working install of Graphite and Carbon: http://graphite.readthedocs.io/en/latest/install-virtualenv.html

Start by installing biggraphite:

$ # Last release version
$ pip install biggraphite
$ # From git
$ pip install -U https://github.com/criteo/biggraphite/archive/master.zip

Configure Cassandra (you will probably want to tweak the keyspace).

$ ${CASSANDRA_HOME}/bin/cqlsh < share/schema.cql
# Can be slow on HDD. Restart-it if it timeouts.
$ bgutil --cassandra_contact_points=localhost syncdb --storage-schemas storage-schemas.conf
# To create arbitrary retentions and see the schema:
$ bgutil --cassandra_contact_points=localhost syncdb --dry_run --retention "86400*1s:10080*60s"

Carbon plugin

In carbon.conf set the following settings:

BG_DRIVER = cassandra
BG_CASSANDRA_KEYSPACE = biggraphite
BG_CASSANDRA_CONTACT_POINTS = 127.0.0.1
DATABASE = biggraphite

You can optionally use whisper+biggraphite or biggraphite+whisper to enable double writes.

You'll need to start bg-carbon-cache instead of carbon-cache to run carbon with biggraphite. If carbon is installed in /opt/graphite/lib and not in site-packages you'll need to set do set PYTHONPATH=/opt/graphite/lib.

To use Elasticsearch as a metadata backend, replace BG_DRIVER parameter by BG_DATA_DRIVER and BG_METADATA_DRIVER parameters and provide informations to connect to Elasticsearch:

# Usual configuration
...

# Cassandra configuration
BG_DATA_DRIVER = cassandra
BG_CASSANDRA_KEYSPACE = biggraphite
BG_CASSANDRA_CONTACT_POINTS = 127.0.0.1

## Elasticsearch configuration
BG_METADATA_DRIVER = elasticsearch
BG_ELASTICSEARCH_HOSTS = 127.0.0.1
BG_ELASTICSEARCH_PORT = 9200

If Elasticsearch needs credentials, they can be provided by BG_ELASTICSEARCH_USERNAME and BG_ELASTICSEARCH_PASSWORD either in the configuration file either as environment variables.

Graphite Web plugin

In local_settings.py set the following settings:

STORAGE_FINDERS = ['biggraphite.plugins.graphite.Finder']
TAGDB = 'biggraphite.plugins.tags.BigGraphiteTagDB'
BG_DRIVER = 'cassandra'
BG_CASSANDRA_KEYSPACE = 'biggraphite'
BG_CASSANDRA_CONTACT_POINTS = '127.0.0.1'

You can optionally keep the StandardFinder to enable double reads from whisper and biggraphite.

To use an hybrid backend (Cassandra as data backend and Elasticsearch as metadata backend), use similar configuration as for Carbon plugin.