To begin using ARA, you’ll first need to set up Ansible so it knows about the the ARA callback and, if necessary, the ara_record and ara_read modules.

The callback and modules are bundled when installing ARA but you need to know where they have been installed in order to let Ansible know where they are located.

This location will be different depending on your operating system, how you are installing ARA and whether you are using Python 2 or Python 3.

ARA ships a set of convenience Python modules to help you configure Ansible to use it.

They can be used like so:

$ python -m ara.setup.path

$ python -m ara.setup.action_plugins

$ python -m ara.setup.callback_plugins

$ python -m ara.setup.library

Using ansible.cfg

This sets up a new ansible.cfg file to load the callbacks and modules from the appropriate locations:

$ python -m ara.setup.ansible | tee ansible.cfg

Or alternatively, if you have a customized ansible.cfg file, you can retrieve only what you need using the other helpers such as the following:

  • python -m ara.setup.callback_plugins
  • python -m ara.setup.action_plugins
  • python -m ara.setup.library

Using environment variables

Depending on the context and your use case, configuring Ansible using environment variables instead of an ansible.cfg file might be more convenient.

ARA provides a helper module that prints out the necessary export commands:

$ python -m ara.setup.env
export ANSIBLE_CALLBACK_PLUGINS=/usr/lib/python2.7/site-packages/ara/plugins/callbacks
export ANSIBLE_ACTION_PLUGINS=/usr/lib/python2.7/site-packages/ara/plugins/actions
export ANSIBLE_LIBRARY=/usr/lib/python2.7/site-packages/ara/plugins/modules

Note that the module doesn’t actually run those exports, you’ll want to run them yourself, add them in a bash script or a bashrc, etc.


ARA uses the same mechanism and configuration files as Ansible to retrieve it’s configuration. It comes with sane defaults that can be customized if need be.

The order of priority is the following:

  1. Environment variables
  2. ./ansible.cfg (In the current working directory)
  3. ~/.ansible.cfg (In the home directory)
  4. /etc/ansible/ansible.cfg

When using the ansible.cfg file, the configuration options must be set under the ara namespace, as follows:

variable = value


The callback, CLI client and web application all share the same settings. For example, if you configure the database location, all three will use that location.

Parameters and their defaults

Environment variable [ara] ansible.cfg variable Default value
ARA_DIR dir ~/.ara
ARA_DATABASE database sqlite:///~/.ara/ansible.sqlite
ARA_PORT port 9191
ARA_LOG_CONFIG logconfig None
ARA_LOG_FILE logfile ~/.ara/ara.log
ARA_LOG_FORMAT logformat %(asctime)s - %(levelname)s - %(message)s
ARA_IGNORE_PARAMETERS ignore_parameters extra_vars
ARA_IGNORE_EMPTY_GENERATION ignore_empty_generation True
ARA_IGNORE_MIMETYPE_WARNINGS ignore_mimetype_warnings True
ARA_PLAYBOOK_OVERRIDE playbook_override None
ARA_PLAYBOOK_PER_PAGE playbook_per_page 10
ARA_RESULT_PER_PAGE result_per_page 25
SQLALCHEMY_ECHO sqlalchemy_echo False
SQLALCHEMY_POOL_SIZE sqlalchemy_pool_size None (default managed by flask-sqlalchemy)
SQLALCHEMY_POOL_TIMEOUT sqlalchemy_pool_timeout None (default managed by flask-sqlalchemy)
SQLALCHEMY_POOL_RECYCLE sqlalchemy_pool_recycle None (default managed by flask-sqlalchemy)


Base directory where ARA will store it’s log file and sqlite database, unless specified otherwise.


ARA records Ansible data in a database. The callback, the CLI client and the web application all need to know where that database is located.

ARA ensures the database exists and it’s schema is created when it is run.

ARA comes out of the box with sqlite enabled and no additional setup required. If, for example, you’d like to use MySQL instead, you will need to create a database and it’s credentials:

CREATE USER ara@localhost IDENTIFIED BY 'password';
GRANT ALL PRIVILEGES ON ara.* TO ara@localhost;

And then setup the database connection:

export ARA_DATABASE="mysql+pymysql://ara:password@localhost/ara"
# or
database = mysql+pymysql://ara:password@localhost/ara

When using a different database driver such as MySQL (pymysql), you also need to make sure you install the driver:

# From pypi
pip install pymysql
# For RHEL derivatives
yum install python-PyMySQL
# For Debian or Ubuntu
apt-get install python-pymysql

Alternatively, if you prefer PostgreSQL, you can do the following in psql:


Be sure you update your pg_hba.conf afterwards if needed.

Then, setup the database connection:

export ARA_DATABASE="postgresql+psycopg2://ara:password@localhost:5432/ara"
# or
database = postgresql+psycopg2://ara:password@localhost:5432/ara

You will need to install the database driver by:

# From pypi
pip install psycopg2
# For RHEL derivatives
yum install python-psycopg2
# For Debian or Ubuntu
apt-get install python-psycopg2


The host on which the development server will bind to by default when using the ara-manage runserver command.

It is equivalent to the -h or --host argument of the ara-manage runserver command.


The port on which the development server will listen on by default when using the ara-manage runserver command.

It is equivalent to the -p or --port argument of the ara-manage runserver command.


Path to a python logging config file.

If the filename ends in .yaml or .yml the file will be loaded as yaml. If the filename ends in .json the file will be loaded as json. The resulting dict for either will be treated as a logging config dict and passed to logging.config.dictConfig.

Otherwise it will be assumed to a logging config file and the path will be passed to logging.config.fileConfig.

If this option is given it superseeds the other individual log options.


Path to the logfile to store ARA logs in.


The loglevel to adjust debug or verbosity.


The log format of the logs.


ARA will, by default, save every parameter and option passed to ansible-playbook (except extra-vars) and make them available as part of your reports.

If, for example, you use extra_vars to send a password or secret variable to your playbooks, it is likely you don’t want this saved in ARA’s database.

This configuration allows you to customize what ARA will and will not save. It is a list, provided by a comma-separated values.


When using ara generate html, whether or not to ignore warnings provided by flask-frozen about endpoints for which the application found no available data.

For example, if you do not use the ara_record module as part of your playbooks, this avoids printing a MissingURLGeneratorWarning because there is no recorded data to render.


When using ara generate html, whether or not to ignore file mimetype warnings provided by flask-frozen.


This configuration is exposed mostly for the purposes of the ara generate html and ara generate junit commands but you can use it as well.

ARA_PLAYBOOK_OVERRIDE will limit the playbooks displayed in the web application to the list of playbook IDs specified. This is expected to be playbook IDs (ex: retrieved through ara playbook list) in a comma-separated list.


This is the amount of playbooks runs shown in a single page in the ARA web interface. The default is 10 but you might want to tweak this number up or down depending on the amount of hosts, tasks and task results contained in your playbooks. This directly influences the weight of the pages that will end up being displayed. Setting this value too high might yield very heavy pages.

Set this parameter to 0 to disable playbook listing pagination entirely.


This is the amount of results shown in a single page in the different data tables such as hosts, plays and tasks of the ARA web interface. The default is 25 but you might want to tweak this number up or down depending on your preference. This has no direct impact on the weight of the page being sent for the reports as these data tables are rendered on the client side.

Set this parameter to 0 to disable pagination for results entirely.

The CLI client and the web application

The CLI client and the web application do not need to be run on the same machine that Ansible is executed from but they do need a database and know it’s location.

Both could query a local sqlite database or a remote MySQL database, for example.