This is an example of a simple Slackbot (called '@mentionbot') that posts a notification to a specific channel any time you're @mentioned anywhere. This bot is useful for when you need a time-ordered list of @mentions to go through at a later point.
Slack users can subscribe/unsubscribe from notifications easily. To receive notifications, add @mentionbot
to a channel you want to be notified in. Then send any message to @mentionbot
to subscribe. To stop getting messages, send a message to @mentionbot
containing the word unsubscribe
.
The example contains a few useful patterns showing you how to build a good Slackbot while taking advantage of a lot of conveniences that Pulumi and the aws
and awsx
packages provide.
- We set up an ApiGateway API to receive push notifications from Slack whenever important events happen.
- Slack has strict requirements on how quickly the push endpoint must respond with
200
notifications before they consider the message as "not received", triggering back-off and resending of those same messages. For this reason, our example does not process Slackevent
messages as they come in. Instead, they are immediately added to an AWS SNS Topic to be processed at a later point in time. This allows the ApiGateway call to return quickly, satisfying Slack's requirements. - Two AWS Lambdas are created naturally using simple JavaScript functions. One function is used to create the Lambda that is called when Slack pushes a notification. The other is used to specify the Lamdba that will process the messages added to the Topic. These JavaScript functions can easily access the other Pulumi resources created, avoiding the need to figure out ways to pass Resource ARNs/IDs/etc. to the Lambdas to ensure they can talk to the right resources. If these resources are swapped out in the future (for example, using RDS instead of DynamoDB, or SQS instead of SNS), Pulumi will make sure that the Lambdas were updated properly.
- Pulumi Secrets provides a simple way to pass important credentials (like your Slack tokens) without having to directly embed them in your application code.
First, we'll set up the Pulumi App. Then, we'll go create and configure a Slack App and Bot to interact with our Pulumi App.
Note: Some values in this example will be different from run to run. These values are indicated with
***
.
$ pulumi stack init mentionbot
$ pulumi config set aws:region us-east-2