Oskar - the Slack satisfaction coach

Description

Oskar is a Slackbot that tracks satisfaction of your team members. Every day it asks people how they're doing. This metric is stored in a database and tracked over time, allowing the team to understand which members are struggling or doing extremely well.

Everyone on your team can ask Oskar for another team member's or the entire team's current status. It is not meant to be a way of comparing people but to surface issues, unblock each other and eliminate isolation (especially in remote teams).

Installing him on Heroku

You can deploy your own copy to Heroku with one click using this button:

See the Heroku documentation for more info about changing the configuration variables after deployment.

Using the bot

Oskar automatically asks you two times a day how you're doing. You can reply to him with a number between 1 to 5, and he will then ask you for feedback to know what's the trigger for your current low or high.

All data is collected in a database and made visible via the dashboard, which can be found at the URL: http://your-oskar-url.com/dashboard (find instruction on how to set a username/password below)

You can send the following commands directly to Oskar:

• How is @member? - Tells you how a specific team member is doing
• How is @channel? - Returns the current status for the whole group

Oskar is not just a Slackbot. He also comes with a dashboard view, allowing you to see visualisations of team status over time. This is visible at /dashboard. If you're hosting on heroku, then this URL is going to be something like http://myherokuappname.herokuapp.com/dashboard. See below for info about restricting access to this dashboard.

The App

Tech stack

• Oskar is build on node.js with express.js.
• It is written in CoffeeScript (such as the node slack client it uses)
• It uses a MongoDB database to store team member feedback
• It (usually, but not only) runs on Heroku

Configuring Oskar

There are two ways of configuring Oskar.

1. Using local configuration You copy the contents of the file config/default.json and create a new file config/local.json with your environment's variables.

2. Using Heroku env variables Use .env.sample to set up your Heroku env variables, either setting them via the command line (as described here) or directly from the Heroku panel.

Here's the config variables you need to define:

• mongo.url (or MONGOLAB_URI for Heroku) defines the url to your MongoDB database (to create a mongoDB on Heroku, go to https://elements.heroku.com/addons/mongolab). This will be automatically generated if you create a MongoLab database as described below ("Setting up Oskar on Heroku") in step 4.
• slack.token (or SLACK_TOKEN for Heroku) is the token of your team's Slackbot (you can create a new Slackbot here: https://yourteam.slack.com/services/new/bot)

Additionally you can disable specific channels or users:

• slack.disabledUsers (or DISABLED_USERS for Heroku) to disable channels that Oskar is part of (you should disable the default channel that Slack added)
• slack.disabledChannels (or DISABLED_CHANNELS for Heroku) to disable users if you want specific people on your team to not receive any Oskar messages at all

By default your dashboard is protected via a simple HTTP auth mechanism. (we'll try to improve this in the future)

• auth.username and auth.password (or AUTH_USERNAME and AUTH_PASSWORD for Heroku) define your login data for the dashboard. Make sure to share those with your team.

See the following instructions if you set up Oskar for the first time.

Development and Contributing

Local environment quickstart

• Download and install nodeJS: https://nodejs.org/download/
• Install Grunt: npm install grunt -g
• Run npm install to install dependencies
• Start the local app using Heroku Foreman, with: foreman start web
• You can reach the site at http://localhost:5000
• Compile & watch Sass files: grunt watch. TODO: Document this better because it actually needs to be done before every contribution.

Unit tests

Oskar is being tested with Mocha and should.js

First, we need to install MongoDB if you don't have it running already. Full instructions are here:

$ brew install mongodb
# Create a data folder to store MongoDB databases
$ sudo mkdir -p /data/db
$ sudo chown $USER /data/db

Now we're running, we can initialise the database:

$ mongod

Run the unit tests for all modules:

$ npm test

To run only a single unit test call the test file explicitly, such as npm test test/inputHelper.coffee

For the mongo tests to pass, you'll have to run a mongo database under mongodb://localhost:27017.

You will need to modify the test parameters in package.json under scripts.test, to give the test suite a valid Slack API key to test with. We will shortly be updating the repository so that tests can be run on a Slack test account, though.