search

looker / lookerbot

Lookerbot lets you access all your Looker data from Slack! Super fun!
https://looker.com/lookerbot
MIT License
169 stars 141 forks source link
chat chatbot data data-visualization looker slack slash-commands

readme

Lookerbot

Lookerbot integrates Slack and Looker to put all your data at your fingertips.

With Lookerbot, everyone in your company can easily share data and answer questions instantly. Lookerbot can answer questions, send alerts, and more!

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

Features

Detailed information on how to interact with Lookerbot can be found in the Looker Help Center.

Requirements

Deployment

Create a new bot in Slack

  1. Go to https://api.slack.com/apps?new_classic_app=1 and create a new Classic Slack App. Lookerbot uses the Real Time Messaging API so it must be a Classic Slack App. We call the app Looker but it's up to you.
  2. Click on the "OAuth and Permissions" tab and in the Scopes section add the following scopes using the "Add An OAuth Scope" button - don't click the green "Update Scopes" button, as this will convert the app to the new scopes workflow and will break your app! :
    1. channels:read
    2. chat:write:bot
    3. files:write:user
    4. team:read
    5. users:read
    6. commands (if you plan on configuring slash commands)
  3. Navigate to "App Home" and click "Add Legacy Bot User"
  4. At the top of the "OAuth & Permissions" page, click "Install App to Workspace."
  5. Click "Allow" to allow the Slack workspace to use the newly created app.
  6. At the top of the "OAuth & Permissions" page, copy the "Bot User OAuth Access Token" (you'll need this later). Note: bot token should start with xoxb-.
  7. Under "Basic Information", you can add an icon and description for Lookerbot. Here's the icon we use.

    By default, Slack Apps are internal to your team. Don't "distribute" your Slack App – that will make it available to all Slack users in the world.

[!IMPORTANT] Please note: some of the Environment Variables below have changed. You may need to adjust them in order to keep this working.

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 prompt you to give the app a unique name, add the Slack API key and configure all of the required variables (see "Environment Variables" below).

Once the environment variables have been set and the server has been deployed, the bot should be ready to go! You can also optionally configure slash commands.

Troubleshooting

See dependency issues on heroku? Apply YARN_PRODUCTION=false as env. to the deployment. See heroku skip-pruning for more details.

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.

Environment Variables

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

If you'd like to put these configuration variables on the filesystem instead, you can place them in a .env file at the root of the project as well. Environment variables will take precedence over .env settings if both are present.

Tweaking Behavior

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

(optional) Storage Services for Visualization Images
Amazon S3
Microsoft Azure
Google Cloud Storage

If Lookerbot is running on Google Compute Engine, no further information should be needed if the approprate API scopes are set up.

Otherwise, you can provide credentials directly:

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 thoroughly 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:

[{"url": "https://me.looker.com", "apiBaseUrl": "https://me.looker.com:19999/api/4.0", "clientId": "abcdefghjkl", "clientSecret": "abcdefghjkl"},{"url": "https://me-staging.looker.com", "apiBaseUrl": "https://me-staging.looker.com:19999/api/4.0", "clientId": "abcdefghjkl", "clientSecret": "abcdefghjkl"}]

The LOOKER_URL, LOOKER_API_BASE_URL, LOOKER_API_CLIENT_ID, LOOKER_API_CLIENT_SECRET, LOOKER_WEBHOOK_TOKEN, and LOOKER_CUSTOM_COMMAND_FOLDER_ID variables are ignored when LOOKERS is set.

Running the Server

To run the server:

  1. Ensure Node.js is installed
  2. yarn install to install dependencies
  3. yarn 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. Go to https://api.slack.com/apps and find your app.
  2. Choose "Slash Commands" and click "Create New Command".
  3. Create a command to use for the Looker bot. We use /looker but it's up to you.
  4. Set the URL to wherever you have your bot server hosted (if you used Heroku to set up the server, this will be the unique app name that you chose) . The path to the slash command endpoint is /slack/receive, so if your server is at https://example.com, the URL would be https://example.com/slack/receive.
  5. Under settings, choose "Install App" again, then "Reinstall App" and authenticate.
  6. Under "Basic Information", grab the verification token. You'll use this to set the SLACK_SLASH_COMMAND_TOKEN environment variable.

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 default. The format selection is ignored.

  4. Enter the webhook URL of the server you set up.

    • Post to public channels /slack/post/channel/my-channel-name
    • (Lookerbot will need to be invited to this channel to post in it.)
    • Post to private groups /slack/post/group/my-channel-name
    • (Lookerbot will need to be invited to this group to post in it.)
    • To direct message a user /slack/post/dm/myusername

    These URLs are prefixed with the URL of your server. (If you used the Heroku deployment, this will be the unique app name you chose). So, if your server is 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.

  5. 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.

Data Actions

Performing Data Actions from Slack

By default, simple data actions will appear in Slack for single value visualizations. Data actions that have forms are not currently supported.

This can be disabled on a per-action basis by using Liquid templating in the action definition to restrict access to certain users. Alternately, the action buttons can be disabled entirely with the bot configuration variable LOOKERBOT_DATA_ACTIONS_IN_MESSAGES.

There's a quick additional configuration that's needed to use Data Actions from Slack:

  1. Go to https://api.slack.com/apps and find your app.
  2. Choose "Interactive Messages" and enable that feature.
  3. For the "Request URL", set the URL to wherever you have your bot server hosted (if you used Heroku to set up the server, this will be the unique app name that you chose). The path to for interactive message requests is /slack/action, so if your server is at https://example.com, the Request URL would be https://example.com/slack/action.
  4. Configure the Slash Command Token as described here.

Sending Slack Messages via Data Actions

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.

dimension: value {
  sql: CONCAT(${first_name}, ' ', ${last_name}) ;;

  # Let user choose a Slack channel to send to
  action: {
    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
  action: {
    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
  action: {
    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 Lookerbot, and using that user's API credentials. It's worth remembering that everyone who can talk to your Lookerbot 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.

Running Locally for Development

  1. Install Node.js on your local machine.
  2. Install Yarn on your local machine.
  3. Add your environment variables to a file called .env at the base of the repo.
  4. Install dependencies with yarn install
  5. Run the bot with yarn 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!