Cloud Services Availability

Query clouds services, save their availability data and provide aggregated information via RESTful API

Configuration

Default location of configuration file is /etc/oss/availability/config.json.

If there is an environment variable AVAILABILITY_CONF set, it is used as priority. In this case it is important to use absolute path to configuration file.

Example:

.. code-block::

export AVAILABILITY_CONF=/home/username/oss_configs/availability.json
python availability/main.py

Configuration File Description

Configuration file is plain JSON document which defines Flask <http://flask.pocoo.org>_, connection to data storage and services quering configuration:

Here is a simple example:

.. code-block::

{ "flask": { "PORT": 5020, "HOST": "0.0.0.0", "DEBUG": false }, "backend": { "type": "elastic", "connection": [{"host": "127.0.0.1", "port": 9200}] }, "period": 60, "regions": [ { "name": "west-1", "services": [ {"name": "nova", "url": "http://foo.example.org:1234/"}, {"name": "cinder", "url": "http://foo.example.org:4567/"} ] }, { "name": "west-2", "services": [ {"name": "keystone", "url": "http://example.org/ab/"}, {"name": "cinder", "url": "http://example.org/cd/"} ] } ], "connection_timeout": 1, "read_timeout": 10, "logging": { "level": "ERROR" } }

flask

Flask configuration is set via *flask* key and described in
`official documentation <http://flask.pocoo.org/docs/0.11/config/>`_.

The only extra options are *HOST* and *PORT*.

backend

Type of backend (currently only "elastic" is supported). Connection to backend (a list of dicts with host and port keys of Elasticsearch <https://github.com/elastic/elasticsearch>_)

period

Period of services availability check, in seconds. Numeric.

regions

List of services to check, by regions. Each item is an object with the following properties:

• name - region name

• services - list of services:

  • name - service name
  • url - service URL to check

connection_timeout

Number of seconds to wait until connection is established

read_timeout

Number of seconds for reading connection response

logging

Logging configuration:

* *level* - set specific logging level

Build Docker Image

To build the image run the following command:

.. code-block:: sh

docker build -t availability:latest .

Run Docker Containers

Edit configuration file in-place in the repository:

.. code-block:: sh

    cd ~/availability/
    cp etc/sample_config.json etc/config.json
    vim etc/config.json

Then run two separate containers with applications:

    docker run -d --name avail-watcher -v $PWD/etc:/etc/availability availability availability-watcher
    docker run -d --name avail-api -v $PWD/etc:/etc/availability -p 5010:5000 availability ./entrypoint-api.sh

Or one all-in-one:

    docker run -d --name avail -v $PWD/etc:/etc/availability -p 5010:5000 availability ./entrypoint-all.sh