Serverless backend example with couchbase

This project is a simple CRUD serverless backend written in nodejs that uses couchbase as database.

Prerequisites

In order to run the project in your environment you need:

• A couchbase cluster (preferably tls enabled):
  • ca.pem (if tls enabled);
  • the cluster connection string;
  • a bucket;
  • the credentials to access the bucket.
• An AWS account.

In order to compile and deploy this project you need:

• Node.js;
• Docker;
• Serverless: configured with your aws credentials.

Project structure

The project is structured as following:

• layers: contains the configuration files needed to build and deploy a lambda layer with couchbase node sdk v2.6.11;
• secrets: contains the CA used to connect securely to the couchbase cluster;
• src: contains the lambdas.

The .env and the ca.pem are crypted with git secret.

Environment variables

The following variables are required:

• CLUSTER_USERNAME;
• CLUSTER_PASSWORD;
• CLUSTER_CONNECTION_STRING;
• BUCKET.

With the serverless-dotenv-plugin all of the environmnet variables will be loaded with the .env in the project root.

Lambda layer

In this example we're using Lambda Layers to store the couchbase sdk.

Why?

The couchbase node sdk is not a pure js module, in fact it uses the underlying C SDK. In order to properly setup the sdk for the lambda we need to run npm install on an amazon-linux OS like. For this purpose we can use Docker to install the node_modules and serverless to store them in a lambda layer. In this way we can continue to change and deploy the lambda code without building each time the couchbase sdk.

How?

Under the layers folder we can find the serverless.yaml, which defines the couchbase-node-sdk-2-6-11 layer. This layer loads the couchbase folder. It's mandatory to run npm run-script build and npm run-script install-couchbase inside the couchbase folder, in order to build the docker image and run a container. Once run, you must see the nodejs/node_modules folder full with the couchbase dependencies. In order to deploy the lambda layer, just run the serverless deploy on the layers folder.

Deploy

Once you have deployed the lambda layer, you can deploy the backend. Be sure to install all the node dependencies with npm install on the project root, than you can run the serverless deploy.

CRUD APIs

Once the deployed the backend consists of 4 APIs:

• Create a new contact with POST /:
  • body: { "name": "", "surname": "", "email": ""};
• Retrieve a contact with GET /{contactId};
• Update a contact with PUT /{contactId};
• Delete a contact with DELETE /{contactId}.

The contact model has the following attributes:

• name: required, at most 64 chars;
• surname: required, at most 64 chars;
• email: optional, valid email with at most 64 chars.