contrail 0.3.0

Public cloud market price tracker for Amazon EC2 and Azure

Contrail

Public Cloud Market Price Tracker

UC Davis ECS 193, 2019

Contrail is a public cloud market price tracker that allows you to easily access and compare the prices and characteristics of VMs over various locations, cloud providers, and times.

Quick Start

Virtual Environment

Create virtual environment:

virtualenv -p python3 venv
OR
python3 -m virtualenv venv

Start virtual environment:

source venv/bin/activate

You should see (venv) show up in command prompt.

Install requirements our project uses:

pip install -r requirements.txt

To exit virtual environment:

deactivate

Configure

Find config_example.py in the main Contrail folder and make a copy named config.py. For a quick start, change the line

CLICKHOUSE_DB_URL = 'http://localhost:8123'

to

CLICKHOUSE_DB_URL = 'http://contrail.tk:8123'

This will make the frontend pull data from the official Contrail database. To collect your own data, see Collecting your own data below.

Start Django

Make sure you are in the main Contrail folder and the virtual environment is activated.

Collect static files:

python contrail.py frontend collectstatic

Start the Django server:

python contrail.py frontend runserver

Open 127.0.0.1:8000 in a browser.

API

/api/getinstances/
Given attributes, return instances that match and their prices
request:

{
    'operating_system': "Linux",
    'aws': "True",
    'gcp': "True",
    'azure': "False",
    'region': None,
    'vcpus': None,
    'memory': 8,
}

response:

{
    "instance_type": "c4"
    "operating_system": "Linux",
    "provider": "AWS",
    "region": "US East",
    "vcpus": 8,
    "memory": 8,
    "reserved", "spot",
    "price_type": "on_demand",
    "price": 0.233,
    "price_unit": "per hour"
}

Collecting your own data

Contrail is split into a few distinct components:

• Crawler: Collects and stores raw offer data from various cloud providers.
• Loader: Parses raw offer data and stores it in a Clickhouse data warehouse.
• API/Frontend: Communicates with the data warehouse and displays data nicely.

The Quick Start above only sets up the API/Frontend component. By default, it points to our own data source. You will have to set up your own backend (crawler and loader) to collect your own data.

Read more about setting up the first two components here: Crawler, Loader.