Summary

Get up an running to develop with RStreams with the sample app in less than 30 minutes.

Install Node and NPM

Install Node and NPM. RStreams Flow supports whatever version of Node JS is deployable to AWS lambda. At the time of this article, that’s Node 14.

Install the Serverless Framework

Install the Serverless Framework as a global package on your development box.

npm install -g serverless

Get bus stack name

You must have an RStreams Bus instance to access. Each bus instance installs a secret in AWS secrets manager named rstreams-{busName}. The bus name is something very specific. The RStreams Bus stack itself installs four child stacks. One of those child stacks is the actual bus. So, if I have a bus installation named PlaygroundBus then there will be a stack named PlaygroundBus-Bus-<random-characters-AWS-puts-on>. You need this name. Go to AWS CloudFormation and click on Stacks and then in the search field search on the name of your bus. In my case, I searched on the word playground and the five stacks showed up the main stack and the five child stacks. I found the exact name of the bus stack which was PlaygroundBus-Bus-1JX7JSIIUQRAO.

Be sure you are using a test bus instance. Many companies will install a bus in their dev or test environments/accounts. If you’re unsure, reach out to IT/Devops in your company. You can also standup your own bus instance but be warned this creates some AWS resources: Kinesis, S3, a few DynamoDB tables, a secret in Secrets Manager. It’s not much money, around 50 to 100 bucks a month, to just stand one up and use it lightly for testing but if you push lots of data through your bus, it will become real money. Be sure to contact your IT/Devops/Engineering management on their policies around this.

AWS Access

You are going to need to access from your laptop to the resources listed in the AWS Secrets Manager secret named for the RStreams bus instance you wish to use. Those resources are six DynamoDB tables, a Kinesis stream, a Firehose stream and a private S3 bucket. Please reach out to your IT/Devops/Eng mgt to follow your company’s proscribed policy for gaining access to these resources. If IT/Devops/Eng mgt is just you, there are lots of online resources on creating IAM policies. Here are the contents of rstreams-TestBus that might exist in a secret for say a test environment.

 1{
 2  // These are DynamoDB tables
 3  "LeoStream":"TestBus-LeoStream-1234567",
 4  "LeoCron":"TestBus-LeoCron-1234567",
 5  "LeoSettings":"TestBus-LeoSettings-1234567",
 6  "LeoEvent":"TestBus-LeoEvent-1234567",
 7  "LeoSystem":"TestBus-LeoSystem-1234567",
 8
 9  // A Kinesis stream
10  "LeoKinesisStream":"TestBus-LeoKinesisStream-1234567",
11
12  // A Kinesis Firehose stream
13  "LeoFirehoseStream":"TestBus-LeoFirehoseStream-1234567",
14
15  // An S3 bucket
16  "LeoS3":"testbus-leos3-1234567",
17
18  // The region this is installed in
19  "Region":"us-east-1"
20}

Create a new project from an RStreams Flow Template

Step 1 - Create the project from the template

1. Drop to the command line and CD to where you will want your new project to be
2. Think of a name for your project: lowercase and separate words with dashes, no spaces and no other punctuation. If this is just a test project for you, name it something like test-123131. Be sure that this name is unique in case others at your company are using this guide, maybe add your initials or something. If this will be the basis of a real project, name it for the microservice you are creating. What matters is that the name is unique at your company. The reason for this is that by convention, this sample project will name bots and queues prepended with your project name, which is used as your service name. Of course, you can change this later as you understand how things work more.
3. Run this command to create a new project, replace {your-project-name} with the name you thought of in the previous step.

serverless create --template-url https://github.com/LeoPlatform/rstreams-flow-example/tree/master -p {your-project-name}

Step 2 - Initialize everything

1. CD to the root of your project
2. run npm install
3. run this to initialize everything using your project’s name as the service name for bots/queues

   serverless init-template
   

4. You will be prompted for both the region your RStreams Bus is installed within and the rstreams bus stack name, provide them - here’s what the output looked like

   

Step 3 - Compile and Run

1. Run tsc or npx tsc from the command line to compile TS to JS (execute npm run watch to have a watcher compile TS to JS when files change)
2. Run the weather-loader bot locally hitting an actual RStreams queue

npm test weather-loader

YOU ARE UP AND RUNNING!

Project Commands using NPM

npm run coverage-all

npm run coverage

npm run utest

npm run package

npm run deploy-dev

npm run test-sls
npm run test

npm run watch

npm run watch-config

npm run webpack

npm run lint