GitHub Comment-Triggered commands

The gh-comments source plugin showcases how to trigger Botkube executors from GitHub comment.

The posted comment on the watched repository is parsed; if it matches the prefix, the command is executed, and the output is posted on the configured communication platform channel. In this case, the output was posted on a Slack channel. However, it can be any communication platform that Botkube supports, even outgoing webhook, e.g.:

Tutorial

Prerequisites

• Kubernetes cluster (e.g. k3d)
  note

  To start a local cluster, run k3d cluster create

Steps

1. (optional) Obtain the GitHub access token. Instruction for creating a token can be found here: https://help.github.com/articles/creating-a-personal-access-token-for-the-command-line/#creating-a-token. This is used to increase the rate limit or watch private repositories. If you don't want to use it, leave the accessToken field empty. The API that is called is described here.

2. Create the config.yaml configuration file, with the following content:

   sources:
     "github":
       displayName: "GitHub comment watcher"
       mszostok/gh-comments:
         enabled: true
         config:
           github:
             auth:
               # access token is optional, it is used to increase the rate limit or watch private repositories
               accessToken: ""
           onRepository:
             # The name of the GitHub repository in the form 'owner/repository'.
             # required
             name: "mszostok/gh-test-repo"
             # the interval for rechecking for a new comments on configured repository.
             # default: 5s
             recheckInterval: 5s
             # The interval for rechecking for a new comments on configured repository. If empty, all comments will be processed.
             # default: "#run"
             commentRequiredPrefix: "#run"
   
   actions:
     "run-commands":
       enabled: true
       displayName: "Run commands from incoming source"
       command: "{{ .Event.command }}"
       bindings:
         sources:
           - github
         executors:
           - k8s-default-tools
   
   executors:
     k8s-default-tools:
       botkube/kubectl:
         enabled: true
       botkube/helm:
         enabled: true
   settings:
     clusterName: "labs"
   plugins:
     repositories:
       mszostok:
         url: https://github.com/mszostok/botkube-plugins/releases/download/v1.2.0/plugins-index.yaml

3. Install Botkube with the config.yaml configuration file and configured communication platform.

Webhook

helm repo add botkube https://charts.botkube.io
helm repo update

export WEBHOOK_URL={url} # e.g. obtain from https://webhook.site/

helm install botkube botkube/botkube --version v1.7.0 \
-f /tmp/config.yaml \
--set communications.default-group.webhook.enabled=true \
--set communications.default-group.webhook.url=${WEBHOOK_URL} \
--set communications.default-group.webhook.sources={"github"} \
--set image.repository=kubeshop/pr/botkube \
--set image.tag=1341-PR

Socket Slack

helm repo add botkube https://charts.botkube.io
helm repo update

export SLACK_CHANNEL_NAME={channel_name} # without '#'
export SLACK_API_APP_TOKEN={token} # starts with 'xapp-1'
export SLACK_API_BOT_TOKEN={token} # starts with 'xoxb-'

helm install botkube botkube/botkube --version v1.7.0 \
-f /tmp/config.yaml \
--set communications.default-group.socketSlack.enabled=true \
--set communications.default-group.socketSlack.channels.default.name=${SLACK_CHANNEL_NAME} \
--set communications.default-group.socketSlack.appToken=${SLACK_API_APP_TOKEN} \
--set communications.default-group.socketSlack.botToken=${SLACK_API_BOT_TOKEN} \
--set communications.default-group.socketSlack.channels.default.bindings.sources={"github"} \
--set 'executors.k8s-default-tools.botkube/kubectl.enabled'=true \
--set 'executors.k8s-default-tools.botkube/helm.enabled'=true

For other installation options, see [**Botkube installation**](https://docs.botkube.io/installation) document.

4. Navigate to configured GitHub repository and add a comment with the prefix #run to trigger the command. For example, to run kubectl get pods command, add the following comment #run kubectl get pods to the issue or pull request.