ARC Lambda

This repo manages examples using lerna-nx to create ARC backend services and nodejs functions which can be deployed on AWS lambda functions using cdktf and arc-cdk

Pre-Requisites

• node.js
• npm
• use VS code for best development experience
• Terraform
• cdktf-cli

How to use

• Clone the github repo
• cd into the folder and run npm i to install node_modules (and to create envs).
• Run 'npx lerna bootstrap` to install dependencies in workspace folders.
• packages/* contains all the ARC backend services supported by cdktf in AWS lambda.
• To run any service locally , cd into the service folder and update reqd values in .env file (more info env variables in the service readme.md)
• Run npm start to start the development server

How to deploy

To deploy service on AWS lambda :

• Run npm run build to generate the code build
• Run npm run build:layers to generate the node_modules as lambda layers
• Run npm run build:migrations to build install dependencies for migration code which will be deployed as separate lambda function
• We can choose to skip running commands mentioned in step 1-3 and directly run npm run build:all to build the lambda layers, code build and migrations for the service
• cd into cdk folder inside the service and update the .env file(Make sure upstream dependencies like postgres DB etc are already setup)
• Run npm run deploy to deploy the code on AWS using terraform constructs

Checkout out README files inside each example to know more.

Use Lerna

To generate build for all packages

npx lerna run build:all

To deploy build single package

npx lerna run build:all --scope="service-name"

List of supported services

• arc-audit
• arc-auth
• arc-bpmn
• arc-feature-toggle
• arc-in-mail
• arc-payment
• arc-scheduler
• arc-search
• arc-user-tenant

How to Test

Here's how you can test the deployed lambda(s) for this project:

• Set the LAMBDA_URL variable in the env to the Domain name or the url of the api gateway that triggers the lambda.

• To test all the deployed services:

  • Run lerna run test:lambda. - Please note: If you encounter error inotify_add_watch system call has failed due to insufficient space on the device on your linux system. Try to increase the number of inotify watches. - To see current number of inotify watches. Run the command cat /proc/sys/fs/inotify/max_user_watches in your terminal. - To increase the number of inotify watches run the command sudo sh -c 'echo fs.inotify.max_user_watches=524288 >> /etc/sysctl.conf' sudo sysctl -p

• To test selective services:

  • Run lerna run test:lambda --scope=package_name1 --scope=package-name2 ... --verbose
  • To list all the packages run lerna ls --all

• Adding tests:

  • To add tests, add tests in ${service-name}/src/__tests__/lambda directory.
  • This project uses @loopback/testlab, supertest for writing tests. For more info please refer loopback_testing
  • For hitting an api endpoint you might need to be authorized to complete the action. For authorization you need JWT token with required permissions as Bearer Token. You can generate a JWT using the following nodejs script:

    const jwt = require('jsonwebtoken');
  
    const testUser = {
    id: "id",
    username:"username",
    password: "password",
    permissions: [
    //Add the permissions required to hit an endpoint
    ],
    };
  
    const token = jwt.sign(testUser, 'jwtSecret', {
    expiresIn: 180000,
    issuer: 'sourcefuse',
    });
  

  For more info see jsonwebtoken documentation.

AWS Architecture

The lambdas are designed to run in a private subnet with access to Redis and a compatible database. In the example below, we demonstrate a completely serverless and managed infrastructure AWS stack.

• The Lambdas run in a private subnet, since they need to communicate with backend services such as a database and distributed cache.
• Each Lambda that requires Redis can either provision a standalone ElastiCache Redis cluster (recommended) or reuse a shared instance.
• Resources such as the network, Aurora cluster, Route53 zone, and VPC endpoints are created upstream.

A cross-section of a ARC typical service is shown below.