Serverless Multi-tenant Backend with AWS Lambda

In this tutorial, you will learn how to use Nile’s tenant virtualization from a serverless application, using a todo list application example. We’ll use Drizzle as the ORM to interact with the database, Express as the web framework, NodeJS as the runtime and Serverless Framework for deployment.

1. Create a database

1. Sign up for an invite to Nile if you don’t have one already
2. You should see a welcome message. Click on “Lets get started”
3. Give your workspace and database names, or you can accept the default auto-generated names. In order to complete this quickstart in a browser, make sure you select to “Use Token in Browser”.

2. Create a table

After you created a database, you will land in Nile’s query editor. Since our application requires a table for storing all the “todos” this is a good time to create one:

  CREATE TABLE IF NOT EXISTS "todos" (
    "id" uuid DEFAULT gen_random_uuid(),
    "tenant_id" uuid,
    "title" varchar(256),
    "complete" boolean,
    CONSTRAINT todos_tenant_id_id PRIMARY KEY("tenant_id","id")
  );

You will see the new table in the panel on the left side of the screen, and you can expand it to view the columns.

See the tenant_id column? By specifying this column, You are making the table tenant aware. The rows in it will belong to specific tenants. If you leave it out, the table is considered shared, more on this later.

3. Get credentials

In the left-hand menu, click on “Settings” and then select “Connection”.

Click on the Postgres button, then click “Generate Credentials” on the top right corner. Copy the connection string - it should now contain the credentials we just generated.

4. Set the environment

Enough GUI for now. Let’s get to some code.

If you haven’t cloned this repository yet, now will be an excellent time to do so.

git clone https://github.com/niledatabase/niledatabase
cd niledatabase/examples/serverless/lambda-drizzle

Rename .env.example to .env, and update it with the connection string you just copied from Nile Console. Make sure you don’t include the word “psql”. It should look something like this:

DATABASE_URL=postgres://018b778a-30df-7cdd-b55c-2f9664db39f3:[e-mail]:5432/gwen_db

Install dependencies:

npm install

Optional: You can select a region for deploying this example by editing serverless.yml and changing the region property.

And if you haven’t yet, install the Serverless Framework: npm install -g serverless

5. Deployment

In order to deploy the example, run the following command:

serverless deploy

After running deploy, you should see output similar to:

Deploying serverless-node-drizzle to stage dev (us-east-2)

✔ Service deployed to stack serverless-node-drizzle-dev (93s)

endpoint: ANY - https://z2fmc4ux34.execute-api.us-west-2.amazonaws.com
functions:
  api: serverless-node-drizzle-dev-api (424 kB)

Now you can use curl to explore the APIs. Here are a few examples:

# create a tenant
curl --location --request POST 'localhost:3001/api/tenants' \
--header 'Content-Type: application/json' \
--data-raw '{"name":"my first customer", "id":"108124a5-2e34-418a-9735-b93082e9fbf2"}'

# get tenants
curl  -X GET 'http://localhost:3001/api/tenants'

# create a todo (don't forget to use a read tenant-id in the URL)
curl  -X POST \
  'http://localhost:3001/api/tenants/108124a5-2e34-418a-9735-b93082e9fbf2/todos' \
  --header 'Content-Type: application/json' \
  --data-raw '{"title": "feed the cat", "complete": false}'

# list todos for tenant (don't forget to use a read tenant-id in the URL)
curl  -X GET \
  'http://localhost:3001/api/tenants/108124a5-2e34-418a-9735-b93082e9fbf2/todos'

# list todos for all tenants
curl  -X GET \
  'http://localhost:3001/insecure/all_todos'

6. Check the data in Nile

Go back to the Nile query editor and see the data you created from the app.

SELECT tenants.name, title, complete
FROM todos join tenants on tenants.id = todos.tenant_id;

You should see all the todos you created, and the tenants they belong to.

7. How does it work?

In this section we’ll focus on the serverless aspects. If you want to learn more about how to use Nile with Drizzle, check out our Drizzle getting started guide.

In order to deploy the backend to AWS Lambda, we use the Serverless Framework. The framework handles most of the configuration and deployment for us, and required minimal changes to the code.

The changes we did make:

• Wrapped the Express app with a Serverless handler.
• Made sure the database connections is initialized outside the handler with no top level await
• Added a serverless.yml file to the root of the project. This file contains the configuration for the Serverless Framework.

Lets go over these one by one:

7.1. Wrapping the Express app

In the example, we used serverless-http NodeJS module, which wraps an Express app with a Serverless handler.

Using it is very straightforward:

import serverless from "serverless-http";

const app = express();

// all the application logic goes here - handlers, middleware, etc

export const handler = serverless(app);

7.2. Initializing the database connection

The serverless-http wrapping above is almost enough to get the application running on AWS Lambda. However, we need to make sure the database connection is initialized before the handler is called and remains intact between handler executions.

We are initializing the connection in the db.js file, and exporting the connection object.

const client = new Client({
  connectionString: process.env.DATABASE_URL,
});

export const db = drizzle(client.connect(), { logger: true });

This means that the connection is initialized when the file is imported, during the initial “cold start” of the serverless application, and remains initialized as long as the virtual machine is running. Since in AWS Lambda, the virtual machine is reused between executions, this minimizes the connection overhead for all practical appluications.

7.3. The serverless.yml file

Now we just need to configure the deployment and our serverless application is ready to go. The serverless.yml file contains the configuration for the Serverless Framework and it has 4 important sections.

The first is the general configuration. We set the name of the service, the framework version, and most important - the use of .env file for environment variables.

service: lambda-drizzle
frameworkVersion: "3"
useDotenv: true

Next, there’s the provider section where we configure AWS itself. We set the runtime to NodeJS 18, the region to us-west-2 and include the database connection string in the environment.

The region is important - you want to run your serverless application in the same region as your database. In this case - us-west-2.

provider:
  name: aws
  runtime: nodejs18.x
  region: us-west-2
  environment:
    DATABASE_URL: ${env:DATABASE_URL}

The next section is the functions. Here we connect the API routing to the application handlers. In our case, it is pretty simple - all the routes are handled by the Express app, which then routes them to the correct handler.

functions:
  api:
    handler: app.handler
    events:
      - httpApi: "*"

And last, Serverless Framework has a collection of plugins that extend its basic capabilities. In this case, because we are using Typescript with ES modules, we need to use the serverless-esbuild plugin.

plugins:
  - serverless-esbuild