Don-Bot, the Halo Slackbot

https://store.docker.com/community/images/halotools/don-bot

Summary

This is a chatbot that allows you to query your CloudPassage Halo account without leaving the comfort of your Slack client. No need to log into the web portal to find out the status of a server- just donbot tell me about server XYZ.

Donbot sends messages to the #halo channel by default, or whatever you have specified by the ${SLACK_CHANNEL} environment variable. Only users who are in that channel will be able to interrogate Don-Bot. Requests from any user who is not a member of the #halo channel (or ${SLACK_CHANNEL}, if you overrode the default channel) will be ignored. Consider making that channel private, unless you want everyone in your Slack domain to be able to access Halo through the bot.

It lives in a Docker container, so you can deploy pretty much anywhere. No listening ports, it just establishes a connection to Slack and listens for messages where it's name is mentioned. Then it reaches out to the CloudPassage Halo API to gather information, and drops a report back into the channel where it was requested.

This bot can optionally poll the Halo API events endpoint, and post all critical events into the configured channel. See below for details...

Use a read-only CloudPassage Halo API key...

Running Don-Bot

Requirements:

• CloudPassage Halo READ-ONLY API key
• Slack chatbot token (don-operator.png included in repo for bot profile image)
• An instance of Halo Celery running. (https://hub.docker.com/r/halotools/halocelery)

Doing the thing:

Set the following env vars, and then run:

var	purpose
CELERY_BACKEND_URL	Url for Celery backend
CELERY_BROKER_URL	Url for Celery broker
FLOWER_HOST	Url for Flower host
HALO_API_KEY	Halo API key ID (read-only)
HALO_API_SECRET_KEY	Halo API secret
SLACK_API_TOKEN	Slack token for bot
SLACK_CHANNEL	Notifications go to this channel. Defaults to #halo
MONITOR_EVENTS	Set to yes to send critical events to SLACK_CHANNEL

    docker run -d \
        --name don_bot \
        --restart always \
        -e HALO_API_KEY=$HALO_API_KEY \
        -e HALO_API_SECRET_KEY=$HALO_API_SECRET_KEY \
        -e SLACK_API_TOKEN=$SLACK_API_TOKEN \
        -e SLACK_CHANNEL=$SLACK_CHANNEL \
        -e MONITOR_EVENTS=$MONITOR_EVENTS \
        docker.io/halotools/don-bot

You can add the optional variables, if needed, with: -e OPTIONAL_VAR=$OPTIONAL_VAR

To create a bot user in slack:

• (via Slack) click on Administration -> Manage Apps.

• Search for "bots" in the search box and select the first one.

• Click Add Configuration on the Bot page.

• Choose a username for the bot. "donbot"

• Click on Add bot integration

• The Bot API token will be displayed.

• Invite donbot to a channel, or message it directly.

• donbot help to see available commands.

• Messages picked up by the bot are printed to stdout in the container, which is useful for understanding how users are interacting with it and how it interprets messages.

Extending Don-Bot

• app/donlib/lexicals.py contains Lexicals.get_messsage_type(). That's where the interpretation and extraction happen. If you want to add functionality, that's where you should start.
• There are already some unit tests in app/test/unit/test_unit_lexicals.py to exercise the matching process. That's a great place to start, and take a test-driven approach. Don't even think about offering up a PR for extending Lexicals.get_messsage_type() without having test cases to cover the new work. Unit testing isn't the highest on this project, but app/donlib/lexicals.py is at 100% and needs to stay that way.

Troubleshooting Don-Bot

• donbot health will get you a report of the current availability of all components. If MONITOR_EVENTS is set to yes, you'll also get the timestamp of the last observed event from the API.
• Don-Bot will periodiaclly self-check and if there's an internal component failure, it will attempt to drop a message in-channel and exit. Make sure you start the container with the --restart always argument.
• Failed thread? Use docker logs CONTAINER_NAME to ascertain if there's a stack trace in the bot's logs.
• Not getting the response you expect from interacting with the bot? Have a look at the test cases for lexicals, found in app/test/unit/test_unit_lexicals.py, to see how your statements align with the intended interaction samples in the unit tests.

Author

Feedback goes to [email protected].

License

See LICENSE.txt