Slack bot to automate stuff on cloud.
The bot server receives messages from Slack's Events API. A message exchange (RabbitMQ) is used to send these messages to workers.
You can find details of deployment in the example/deploy folder. This assumes that you have already setup the bot in your Slack's app management page.
The easiest way to install/deploy is with helm charts. First install rabbitmq:
$ helm repo add bitnami https://charts.bitnami.com/bitnami
$ helm install my-release bitnami/rabbitmq
$ cd example/deploy
$ helm install rabbit bitnami/rabbitmq -f rabbitmq.yml
This will create a RabbitMQ service and its deployment. More details about the chart here.
In addition, the environment variables AMQP_SERVICE_HOST and AMQP_SERVICE_PORT are added to bot server and worker pods.
Refer to values.yml for required variables, mainly values required by Slack.
Add Zetta's helm repo and update dependencies to install required sub-charts.
$ helm repo add zetta http://zetta.ai/helm-charts/charts
$ helm repo update
$ helm dependencies update
$ helm install --debug <release-name> .
It is easy to install or deploy just a single worker or a group of related workers as a helm release.
Make sure to add the required environment variables so that workers can connect to the message exchange.
A worker can respond to DMs, channels and threads. Refer to different command types in example/example-workers. Responding to threads is especcially helpful for long tasks and providing updates at checkpoints in a thread, use broadcast=True to notify channel for important checkpoints.
Python package click is used to parse user commands, it provides great support for various use cases.