docker/looker-slackbot

Name: looker-slackbot

Owner: Docker

Description: A super cool Slack bot for Looker!

Forked from: looker/lookerbot

Created: 2017-02-09 18:02:45.0

Updated: 2017-04-24 05:22:40.0

Pushed: 2017-01-23 05:49:44.0

Homepage: http://looker.com

Size: 407

Language: CoffeeScript

GitHub Committers

UserMost Recent Commit# Commits

Other Committers

UserEmailMost Recent Commit# Commits

README

Lookerbot for Slack

Lookerbot for Slack integrates with Looker to allow you to query all of your data directly from Slack. This enables everyone in your company to share data easily and answer data-driven questions instantly. Lookerbot expands Looker URLs in channels and allows you to create custom commands for running saved queries.

For a free trial of Looker go to looker.com/free-trial.

Features

Detailed information on how to interact with Lookerbot can be found on Looker Discourse.

Requirements
Deployment
Create a new bot in Slack
  1. Under “Customize Slack” > “Configure” > “Custom Integrations” select “Bots”
  2. Choose “Add Configuration”
  3. Create a username for your Slack bot. We use @looker but it's up to you.
  4. Choose an icon for the Slack bot. Here's the icon we use.
  5. Grab the API token from the settings page, you'll need this when you set up the bot server.
Heroku Deployment

Deploy

The quickest way to deploy the bot is to use Heroku's one-click deploy button, which will provision a server for your bot. This will also allow you to configure all of the required variables.

Once deployed, the bot should be ready to go! You can also optionally configure slash commands.

Manual Deployment

The bot is a simple Node.js application. The application needs to be able to reach both your Looker instance's API and Slack's API. If you have a self-hosted instance of Looker, be sure to open up port 19999 (or your core_port) in order to accesss the Looker API.

The bot is configured entirely via environment variables. You'll want to set up these variables:

(optional) Amazon S3 Image Storage (optional) Azure Image Storage

If you'd like to put these configurations on the filesystem, you can place them in a .env file at the root of the project and start the bot using node-foreman as described below.

Self-signed or invalid certificates

If your Looker instance uses a self-signed certificate, Lookerbot will refuse to connect to it by default.

Setting the NODE_TLS_REJECT_UNAUTHORIZED environment variable to 0 will instruct Lookerbot to accept connections with invalid certificates. Please ensure you have thouroughly evaluated the security implications of this action for your infrastructure before setting this variable.

This should only impact on-premise deployments of Looker. Do not set this environment variable if Looker hosts your instance.

Connecting the bot to multiple Looker instances

If you would like the bot to connect to multiple instances of Looker, then you can configure the bot with the LOOKERS environment variable. This variable should be JSON array of JSON objects, each representing a Looker instance and its authentication information.

The JSON objects should have the following keys:

Here's an example JSON that connects to two Looker instances:

rl": "https://me.looker.com", "apiBaseUrl": "https://me.looker.com:19999/api/3.0", "clientId": "abcdefghjkl", "clientSecret": "abcdefghjkl"},{"url": "https://me-staging.looker.com", "apiBaseUrl": "https://me-staging.looker.com:19999/api/3.0", "clientId": "abcdefghjkl", "clientSecret": "abcdefghjkl"}]

The LOOKER_URL, LOOKER_API_BASE_URL, LOOKER_API_3_CLIENT_ID, LOOKER_API_3_CLIENT_SECRET, LOOKER_WEBHOOK_TOKEN, and LOOKER_CUSTOM_COMMAND_SPACE_ID variables are ignored when LOOKERS is set.

Running the Server

To run the server:

  1. Ensure Node.js is installed
  2. npm install to install dependencies
  3. npm start to start the bot server. The server will run until you type Ctrl+C to stop it.

The included Procfile will also allow you to run the app using foreman or node-foreman. These libraries also provide easy ways of creating scripts for use with upstart, supervisord, and systemd.

Configuring Slash Commands

Slash commands are not required to interact with the bot. You can DM the bot directly or mention the bot like:

@looker help

and use all the functionality.

However, Slash commands are a bit friendlier to use and allow Slack to auto-complete so you'll probably want to set those up.

  1. Under “Customize Slack” > “Configure” > “Custom Integrations” select “Slash Commands”
  2. Choose “Add Configuration”
  3. Create a command to use for the Looker bot. We use /looker but it's up to you.
  4. Choose an icon for the slash command responses. Here's the icon we use.
  5. Set the URL to wherever you have your bot server hosted. The path to the slash command endpoint is /slack/receive, so if your bot is hosted at https://example.com, the URL would be https://example.com/slack/receive.
  6. You'll need to copy the token provided when you created the slash command and set the SLACK_SLASH_COMMAND_TOKEN variable with it for the bot to accept slash commands.

Directions for creating slash commands can be found in Looker Discourse

Scheduling Data to Slack

You can use the bot to send scheduled Looks to Slack.

  1. Click “Schedule” on a Look

  2. Set “Destination” to “Webhook”

  3. Leave “Format” set to “HTML Attachment”. The format selection is ignored.

  4. Enter the webhook URL.

  5. Post to public channels /slack/post/channel/my-channel-name

    • (Lookerbot will need to be invited to this channel to post in it.)

  6. Post to private groups /slack/post/group/my-channel-name

    • (Lookerbot will need to be invited to this group to post in it.)

  7. To direct message a user /slack/post/dm/myusername

    These URLs are prefixed with the URL your bot. So, if yoru bot is hosted at https://example.com and you want to post to a channel called data-science, the URL would be https://example.com/slack/post/channel/data-science.

  8. You'll need to make sure that the LOOKER_WEBHOOK_TOKEN environment variable is properly set to the same verification token found in the Looker admin panel.

Using Data Actions with Slack

The bot server also implements endpoints to allow you to easily send Data Actions to Slack.

Here's an example of a few data actions you could implement in your LookML. (Replace https://example.com with your bot's hostname.)

To make use of this, you'll need to make sure that the LOOKER_WEBHOOK_TOKEN environment variable is properly set to the same verification token found in the Looker admin panel, just like with scheduling data.

nsion: value {
l: CONCAT(${first_name}, ' ', ${last_name}) ;;

Let user choose a Slack channel to send to
tion: {
label: "Send to Slack Channel"
url: "https://example.com/data_actions"
form_url: "https://example.com/data_actions/form"
param: {
  name: "message"
  value: ":signal_strength: I sent a value from Slack: {{rendered_value}}"
}


Send to a particular Slack channel with a preset message
tion: {
label: "Ping Channel"
url: "https://example.com/data_actions"
param: {
  name: "message"
  value: ":signal_strength: I sent a value from Slack: {{rendered_value}}"
}
param: {
  name: "channel"
  value: "#alerts"
}


Ask the user for a message to send to a particular channel
tion: {
label: "Ask a Question"
url: "https://example.com/data_actions"
form_param: {
  name: "message"
  default: "Something seems wrong... (add details)"
}
param: {
  name: "channel"
  value: "#alerts"
}
Data Access

We suggest creating a Looker API user specifically for the Slack bot, and using that user's API credentials. It's worth remembering that everyone who can talk to your Slack bot has the permissions of this user. If there's data you don't want people to access via Slack, ensure that user cannot access it using Looker's permissioning mechanisms.

Also, keep in mind that when the Looker bot answers questions in Slack the resulting data moves into Slack and is now hosted there. Be sure to carefully consider what data is allowed to leave Looker. Slack retains chat message history on their servers and pushes many types of notifications about messages out via other services.

To allow visualizations to appear in Slack, if configured to do so, the bot uploads them as images to Amazon S3 with an extremely long randomly-generated URL. Anyone with this URL can access that image at any time, though it should be extremely difficult to guess.

If you choose to remove the image files from S3, the Slack messages that relied on those images will be blank.

Tweaking Behavior

There are a couple environment variables that can be used to tweak behavior:

Running Locally for Development
  1. Install Node.js on your local machine.
  2. Install node-foreman with npm install -g foreman
  3. Add your environment variables to a file called .env at the base of the repo.
  4. Install dependencies with npm install
  5. Run the bot with nf start
Contributing

Pull Requests are welcome ? we'd love to have help expanding the bot's functionality.

If you have any trouble with the bot, please open an issue so we can help you out!

This work is supported by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number U24TR002306. This work is solely the responsibility of the creators and does not necessarily represent the official views of the National Institutes of Health.