Introduction

The AWS Ops Wheel is a tool that simulates a random selection from a group participants that weights away from participants recently chosen. For any group, the selection can also be rigged to suggest a particular participant that can be in a blatantly obvious (and sometimes hilarious) way.

Get your own in just a few clicks by starting here:

Or, simply set up a CloudFormation stack using the S3 template url: https://s3-us-west-2.amazonaws.com/aws-ops-wheel/cloudformation-template.yml

We are aware of an issue where you can only run this stack in us-west-2 if you launch from this template, we are working on removing this limitation. This limitation does not apply if you build your own stack using the Development Guide below.

The endpoint will then be in the Cloudformation Stack Output messages.

ScreenShots

Wheels Table

Participants Table

Wheel (pre-spin)

Wheel (post-spin)

User Guide

Concepts

Wheel A group of participants that can be selected from. Users can get a suggestion of a participant from a wheel that is weighted away from recently-chosen participants.

Participant A member of a wheel identified by a name, which must be unique, and also a follow-through url when they are chosen. Participants all start with a weight of 1.0.

Operations

Wheel Operations

• Create a new wheel
• Edit an existing wheel
• Delete a wheel
• Spin the wheel and suggest a participant
  • Notes: This does not adjust weighting, so if you're unhappy with the result, you can spin again.
• Proceed: Accept the suggested participant
• Reset: Restart all participants to equal weights as 1.0

Participant Operations

Notes: Participants aren't shared between wheels

• Add a participant to a wheel
  • This requires a name and url that will be opened in a new browser tab when the participant is chosen. A participant begin with a weight of 1.0 which will always be the average weight for all participants.
• Edit a participant's name and/or url
• Delete a specific participant from the wheel
• Rig a specific participant to be selected next
  • This doesn't change any weighting, but actually bypasses the suggestion algorithm to always suggest the participant until told to proceed.
  • After proceeding, weights are adjusted as if the participant had been selected normally.
  • The rigging can be hidden (deceptive) or non-hidden (comical).

Authentication and User management

AWS Ops Wheel is protected by Amazon Cognito authentication. It uses Cognito User Pools to manage users that have access to the deployed application. By default, during the initial deployment phase, it creates an admin user with a random password that is sent to the email address provided to the run script. During the first attempt to login to the AWS Ops Wheel, the admin user will be asked to change the random password to a new one.

If you need to add more users that have access to the wheel application, you can add them using AWS Cognito web console or using the AWS Cli.

The Weighting algorithm

Assumption: total_weight == number_of_participants == len(wheel). This is because we only redistribute weights among participants and all participants start with a weight of 1.0. The below is the algorithm in python pseudo-code:

def suggest_participant(wheel):
    target_number =  len(wheel) * random()  # Get a random floating point number between 0 and the total_weight
    participant = None
    for participant in wheel:
        target_number -= participant.weight
        if target_number <= 0:
            break
    return participant

def select_participant(chosen, wheel):
    # When there is only one participant in the wheel, the selected participant's weight remains intact.
    # Otherwise, the remaining participant(s) get a slice of the selected participant's weight. That participant will not be chosen on next spin unless it's rigged.
	 if len(wheel) > 1:
	    weight_slice = chosen.weight / (len(wheel) - 1)
	    for participant in wheel:
	        if participant == chosen:
	            participant.weight = 0
	        else:
	            participant.weight += weight_slice

Development Guide

Notes: The development tools are currently only written to be Linux/OSX compatible

Development Dependencies

• NodeJS 6.10+
• Python 3
  • boto3
  • pyaml
  • pytest
  • pytest-cov
  • moto
• AWSCLI 1.11+
• An AWS Account you have administrator privileges on

A dedicated IAM User (Optional, but highly-recommended)

• You should create a dedicated IAM User for AWS Ops Wheel development

Create a custom IAM Policy for the User

• Go to the AWS Create Policy Wizard
• Go to the JSON tab and paste in the content of our policy configuration
• Click Review Policy
• Give it an identifying name (we'll need it for the next step) like AWSOpsWheelDevelopment

Create an IAM user with the policy attached

• Got to the AWS Create User Wizard
• Give it a descriptive name like AWSOpsWheelDevelopmentUser and check the Programmatic access checkbox. Note: It doesn't need to be the same as the name of the policy, but it might help keep your things organized
• Click Next: Permissions
• Switch to the Attach existing policies directly tab and filter on the name you used during the Create custom Policy step
• Click the checkbox next to the policy and click Next: Review
• Click Create user
• On the next page, save the Access key ID and the Secret access key (visible by clicking Show) for use in the AWS Cli Configuration step. Note: This will be the only opportunity to copy the Secret Access Key for this Access Key ID. if you don't copy the secret access key now, you'll need to create a new access-key, secret-key pair for the user.

AWS Cli Configuration

For the purpose of our work, we will use AWS Cli to simplify management of the resources. Later we will add support for the Launch Stack button which will be displayed on the GitHub Repo page.

In $HOME/.aws/config add in your credentials configuration and default region, replacing with your IAM user's credentials (or your own access key and secret key if you didn't follow our highly-recommended best-practice). Note: The region can be whatever region you choose, but you should definitely set a default region. We chose us-west-2 since we're in Seattle and it's close by.

[default]
aws_access_key_id = ACCESS_KEY
aws_secret_access_key = SECRET_KEY
region = us-west-2

Test the code

Currently we have unit tests for the API and the UI. To run the API unit tests, go to the <PATH_TO_YOUR_WORKSPACE>/api directory and run:

pytest --verbose --cov-report term-missing --cov ./ -s

To run the UI unit tests, go to the <PATH_TO_YOUR_WORKSPACE>/ui directory and run:

npm run test

Build and deploy the code

Go to the <PATH_TO_YOUR_WORKSPACE> directory and run:

$ ./run \
  --suffix <SUFFIX, optional with default value as no suffix, so stack name will be 'AWSOpsWheel'> \
  --email <EMAIL_ADDRESS, required only during initial stack creation> \
  --no-clean <CLEAN_BUILD_DIRECTORY, optional with default value as False. Note that do not clean the build directory before building or remove the deploy working directory>

This will:

• Create a ./build directory with all of the build artifacts
• Package the build artifacts up into a zip file with name based on a hash of the contents and upload it to S3 for lambda deployment
• Compile the Service CloudFormation Template:
  • Create the lambda functions for all of the routes in the API
  • Add policies for lambda functions to be called by the gateway's functions
  • Create/update the DynamoDB Tables
  • Create the lambda execution IAM role
  • Create the swagger configuration for API Gateway that points the paths to their functions
• Deploy the template directly to CloudFormation through update or create, depending on if it's a new stack

Start Local Dev Server

Go to the <PATH_TO_YOUR_WORKSPACE>/ui directory and run:

npm run start

Miscellaneous

Import Participant data from .csv file

To populate Participant data from .csv file to one of your wheels you can use a tool that is in utils folder. All parameters are required.

$ <PATH_TO_YOUR_WORKSPACE>/utils/wheel_feeder.py \
  --wheel-url <https://<your_api_gateway>.amazonaws.com> \
  --wheel-id <TARGET_WHEEL_ID> \
  --csv-file-path <PATH_TO_CSV_FILE> \
  --cognito-user-pool-id <COGNITO_USER_POOL_ID> \
  --cognito-client-id <COGNITO_CLIENT_ID>

List Stacks

To list all Stacks that are currently provisioned (or have been in the past):

$ aws cloudformation list-stacks

Delete Stack

To delete existing stack:

$ aws cloudformation delete-stack [--suffix SUFFIX_NAME]

Set up continuous deployment

Create continuous deployment resources:

aws cloudformation create-stack --stack-name AWSOpsWheel --template-body file://cloudformation/continuous-deployment.yml --parameters ParameterKey=AdminEmail,[email protected] --capabilities CAPABILITY_NAMED_IAM

aws cloudformation wait stack-create-complete --stack-name AWSOpsWheel

Make sure you have your preferred CodeCommit access configured.

The following assumes that you are using the AWS CLI Credential Helper.

Push to the newly created git repository:

git config --global credential.helper '!aws codecommit credential-helper $@'

git config --global credential.UseHttpPath true

git remote add app `aws cloudformation describe-stacks --stack-name AWSOpsWheel --query 'Stacks[0].Outputs[?OutputKey==\`RepositoryCloneUrl\`].OutputValue' --output text`

git push app master

Wait for the pipeline to finish deploying:

aws cloudformation describe-stacks --stack-name AWSOpsWheel --query 'Stacks[0].Outputs[?OutputKey==`PipelineConsoleUrl`].OutputValue' --output text

Get the URL of the newly deployed application:

aws cloudformation describe-stacks --stack-name AWSOpsWheel-application --query 'Stacks[0].Outputs[?OutputKey==`Endpoint`].OutputValue' --output text