Launch Go in AWS Lambda - Slack slash command example

There are so many things happening in the 'infrastructure' part of software development. In just few years, we've moved from a grey box under your boss' desk, through virtual machines, things like Chef, Docker, Serverless Architectures.

As much as I don't like buzzwords, for me as a developer, the idea of not having too much concern about where my code is running is quite appealing. I can just focus on 'what' it's doing! Let's see what this 'serverless' means in a nutshell:

Serverless architectures refer to applications that significantly depend on third-party services (knows as Backend as a Service or "BaaS") or on custom code that runs in ephemeral containers (Function as a Service or "FaaS"), the best known vendor host of which currently is AWS Lambda. By using these ideas, and by moving much behavior to the front end, such architectures remove the need for the traditional 'always on' server system sitting behind an application.

So I can run my code without setting up servers, networking etc. and I don't need to pay for the whole thing if it's not used! If you don't have customers yet, you don't need to pay. You don't need to think about fallback servers or how many servers you need to handle rising traffic, you just pay when you use it. That's a huge benefit, especially at the beginning of your road to success. Moreover, that's also great fit for small things that could help your day-to-day life and are often not worth setting up a whole machine if it’s being used once or twice per day. Such characteristic perfectly describes Slack slash commands. In HappyTeam we're using Slack extensively, both for internal communication and for communication with our clients (and if you follow our blog you should also know how to build a Slack bot already).

Now we know what AWS Lambda is. The second ingredient is Go. Although AWS Lambda doesn't support Go yet, I fancy it a lot and try to use it anywhere I can :) So let's see how can we use Go to build a Slack command hosted on AWS Lambda!

Install AWS CLI

AWS CLI will help us communicate with AWS API from terminal and create lambda functions on the fly. Firstly, we need python. I'm on OSX, so I'll just type:

brew install python

in my terminal.
Then we need AWS CLI:

pip install awscli

Now setup AWS CLI with your credentials through:

aws configure

You can check if everything is fine via:

aws s3 ls

If no errors are thrown, we're good to go.

Create role

We’re missing one more thing before rushing to the code - IAM Role for our lambda function. Open up Roles section in AWS console and create a new role named lambda_basic_execution.

In the next step, select AWS Lambda as Role Type:

On Attach Policy screen check AWSLambdaBasicExecutionRole and click Next and then Create Role.
Success!

Create Lambda

Time for some coding. I assume that you have Go installed already (you can find packages and install instructions for your system on official website Downloads – The Go Programming Language, version 1.7.4 is the latest stable at the time of writing this post). Create a new folder in your GOPATH like:

mkdir happy-slash-command
cd happy-slash-command

We will be using GitHub – xlab/go-lambda project, to wrap up our GO code and execute on AWS Lambda (as it's not natively supported yet).

First get the package via:
go get github.com/xlab/go-lambda
Now spin up your favorite code editor and create new file called happy-slash-command.go with following content:

package happyteam

import (
	"bytes"
	"encoding/json"
	"fmt"

	"github.com/xlab/go-lambda/lambda"
)

var Handler = lambda.Use(lambda.HandlerFunc(happyHandler))

func happyHandler(event json.RawMessage, context *lambda.Context) []byte {
	buf := new(bytes.Buffer)
	fmt.Fprintf(buf, "Hello from %s", context.FunctionName)
	return buf.Bytes()
}

It's basically a hello world application for now but you can develop here whatever you want, from simple calculations, through webservice calls to get weather, image processing - you get the idea. For now, we just need something basic, to prove it all works.

Now it's time to use the lib to create lambda function from our code. Open up terminal and execute:

go-lambda create --name "happy-slack-lambda" --role arn:aws:iam::303182444347:role/lambda_basic_execution handler happy-slash-command

In other words, we want to create a lambda function named "happy-slack-lambda" that will use "lambda_basic_execution" role we've created earlier and as its code use handler function from happy-slash-command Go package.

If everything went fine, you should see a summary table in your console, and new lambda function should be visible on your AWS Console:

Go on, click on the lambda name and use blue 'Test' button to see if it works:

Yeey! We've got ourselves a lambda function!

API Gateway

We've got functionality in place, time to make it available outside - we need to set up URL for our lambda.

Get back to your AWS console and open up API Gateway portal. Create new gateway and set its name:

In the configuration panel, create new GET method and set it up to use lambda that we've created (we need to select region and set lambda name).

You should get a similar diagram at this point:

You can test your integration with a blue test button on upper left.

Ok, now from Actions menu, select Deploy API entry and create a new stage:

You should get your URL now!

Looks like it's working!

Onto last step now

Setup Slack Slash command

Open up https:<YOUR_SLACK_TEAM_NAME>.slack.com/apps in your browser and find 'Slash command' integration. On the first screen give it a name, like:

On the next screen you can tweak your slack command, give it a nice icon, change alias etc. For now, we will just paste URL that we've got from AWS API Gateway and set Method to GET:

And we should be good to go here! Open up your Slack and type /happyhi You should see:

Full success!