graphql_aws_pricing_api

AWS Pricing API. The idea was to use this to power terraform-cashier (my next project).

Usage of public endpoint

This code is being run on lambda and can be accessed via the following endpoint: https://fvaexi95f8.execute-api.us-east-1.amazonaws.com/Dev/graphql/

Because of the warmup behavior of Lambda, please give the first call some time (usually around 10-15 seconds) to warm-up.

Usage

An example query string might look like this

{
  AmazonEC2(TermType:"OnDemand", Location:"US East (N. Virginia)", OS:"Linux", InstanceType:"m3.medium", Tenancy:"Shared", PreInstalledSW:"NA",CapacityStatus:"Used") {
    PricePerUnit
  }
}

The SQL equivalent would be

SELECT `PricePerUnit`
FROM   `AmazonEC2`
WHERE  `TermType` = 'OnDemand'
       AND `Location` = 'US East (N. Virginia)'
       AND `InstanceType` = 'm3.medium'
       AND `Tenancy` = 'Shared'
       AND `OS` = 'Linux'
       AND `PreInstalledSW` = 'NA'
       AND `CapacityStatus` = 'Used';

A full example HTTP POST request might look like this

{
  "query": "{\n  AmazonEC2(TermType:\"OnDemand\", Location:\"US East (N. Virginia)\", OS:\"Linux\", InstanceType:\"m3.medium\", Tenancy:\"Shared\", PreInstalledSW:\"NA\") {\n    PricePerUnit\n  }\n}",
  "variables": null,
  "operationName": null
}

An example response might look like this

{
  "data": {
    "AmazonEC2": [
      {
        "PricePerUnit": "0.067"
      }
    ]
  }
}

Aliases

GraphQL Aliases can be taken advantage of to retreive multiple pieces of pricing data at once. For example:

{
  t2_xlarge_Shared: AmazonEC2(TermType:"OnDemand", Location:"US East (N. Virginia)", OS:"Linux", InstanceType:"t2.xlarge", Tenancy:"Shared", PreInstalledSW:"NA",CapacityStatus:"Used") {
    PricePerUnit
  }
  t2_medium_Shared: AmazonEC2(TermType:"OnDemand", Location:"US East (N. Virginia)", OS:"Linux", InstanceType:"t2.medium", Tenancy:"Shared", PreInstalledSW:"NA",CapacityStatus:"Used") {
    PricePerUnit
  }
}

The response for the above query should look like this:

{
    "data": {
        "t2_xlarge_Shared": [
            {
                "PricePerUnit": "0.2208",
                "Unit": "Hrs",
                "Currency": "USD"
            }
        ],
        "t2_medium_Shared": [
            {
                "PricePerUnit": "0.0552",
                "Unit": "Hrs",
                "Currency": "USD"
            }
        ]
    }
}

Here's a working curl you can use to test the above request query

curl -H "Content-Type: application/json" -X POST -d '{"query":"{\n  t2_xlarge_Shared: AmazonEC2(TermType:\"OnDemand\", Location:\"US East (N. Virginia)\", OS:\"Linux\", InstanceType:\"t2.xlarge\", Tenancy:\"Shared\", PreInstalledSW:\"NA\",CapacityStatus:\"Used\") {\n    PricePerUnit\n  }\n  t2_medium_Shared: AmazonEC2(TermType:\"OnDemand\", Location:\"US East (N. Virginia)\", OS:\"Linux\", InstanceType:\"t2.medium\", Tenancy:\"Shared\", PreInstalledSW:\"NA\",CapacityStatus:\"Used\") {\n    PricePerUnit\n  }\n}","variables":null}' https://fvaexi95f8.execute-api.us-east-1.amazonaws.com/Dev/graphql/

Local Development

Pre-Requisites

In order to run this API, you must first have a MariaDB instance that has been populated using this python script. If you have not done so already, please run this script pointing at your MariaDB instance to populate it.

Environment Variables

Variable Name	Description
MARIADB_HOST	The hostname/ip of your MariaDB Instance (e.g. localhost)
MARIADB_USER	The user with which to authenticate to MariaDB
MARIADB_PASSWORD	The password with which to authenticate to MariaDB
MARIADB_DB	The DB name to connect to (e.g. aws_prices)

Instructions

To run this API locally, make sure you have all environment variables set and have a target MariaDB database running.

To install any required dependencies please run npm install

Then, simply start the application using npm start

GraphQL Data Types

See SCHEMATA.md

Local Development with Docker

You can use docker-compose to create a local test environment that should match the deployed environment for the purposes of development and local testing.

To get started, run docker-compose up. This should create the necessary containers for:

• The database
• The graphqsl server
• The data importer

More specifically you might want to do docker-compose up -d --build.

Once the database is running the importer script should initialize the database schema and data.

You can connect to the mariadb cli via:

docker-compose run db mysql -h db -p'prices123'

You can do an data import update via:

docker-compose run importer /scripts/pricing_import.py

You should be able to access the server from http://localhost:4000. Note that the server will not start up properly until the initial data ingestion, so if you are running docker-compose up for the first time you may have to wait a few minutes. Check the status via docker-compose logs server. If it outputs the message that Graphql is running at localhost:4000 you're good to go.

Any updates that you make to the server/ directory will automatically propagate to the container and the app will be restarted by pm2.

TODO

Improve lambda development, perhaps use serverless framework