LAA-Apply Slack bot

A bot designed to allows slack users to run commands to assist with maintaining the LAA Apply applications

Usage

The bot can be invited to a channel /invite @apply-bot or interacted with via direct messages.

Interacting

Should be addressed as @apply-bot

e.g. @apply-bot <command>

Commands

• add users - @apply-bot add users name1, name.2, name.three@provider.com
• ages - @apply-bot ages
• details - @apply-bot cfe details staging or @apply-bot apply details live
• run tests - @applybot run tests
• uat url - @applybot uat url ap-1234
• uat urls - @applybot uat urls
• help - @apply-bot help provides better examples than these

Developing locally

This can be a handful...

The corporate slack instance doesn't allow unsanctioned bots to run. Therefore you will need access to a standalone slack instance.

I have set up a personal, free, slack instance and configured a slack token of my own. See the Slack API documents for help with that. To access all of the commands, you will need to be running redis (redis-server), sidekiq (bundle exec sidekiq -r ./app.rb) and the server itself (```bundle exec rackup -r ./app.rb``)

As an alternative you can use foreman start which will run all three services in a single window, but that is not as helpful for debugging.
Binding.pry breakpoints tend to freeze and/or get lost in the threaded services as they run. YMMV

As interactive elements have been developed you will now need to run a service that allows passing a url to the slack instance e.g. ngrok [Note: I used ngrok but do not necessarily endorse it for all users, use it at your own risk!]

Set this up and start it running on your machine. Then, when it's running and pointing at your local 9292 port (e.g. ngrok http 9292), it will give you a url (e.g. https://1234-56-7-250-215.ngrok.io) that you can add to your Slack instance api/application Interactivity & shortcuts section with the suffix /interactive (e.g. https://1234-56-7-250-215.ngrok.io/interactive) this will allow callbacks from slack button clicks to your local test server while ngrok runs

Deploying

This is handled via github actions - on a pull request being merged to master, rspec and rubocop run.

As long as they are successful, a deploy job runs that will build a new docker container, push it to ECR and then apply that docker tag to the K8s cluster

After approx ~3 minutes the pods should have updated and the bot should respond to hi or your new command