Improving your API response time - A guide to placing email functions on queues

Introduction

Developers are always very curious about how to make API responses as fast as possible even with asynchronous functions, a starting point will be to seperate some jobs to be done at a later time while still giving users quick feedbacks. Starting with email jobs which is most time handled by thrid party/external services, API responses can get slow therefore increasing the total response time.
So we can utilize the power of Queues and Redis to seperate the jobs and monitor the status of such jobs compared to making the requests synchronous with your API's response. An added benefit is the ability to retry a job after it has failed and setup event listeners for each of the jobs.

Prerequisites

To follow along with this tutorial you will need the following:

• NodeJs and Yarn installed on your machine.
• Docker installed on your local machine.
• Basic knowledge of Typescript and NodeJS.

Table of Contents

1. Setup NodeJS Express Project with Typescript and Bull.
2. Setup Redis Server
3. Setting up Queues and Processes
4. Configuring adapters to monitor queues
5. Connecting to a free redis database

Setup NodeJS Express Project with Typescript and Bull

We would be creating a new project and installing all the dependencies in this section.
Let's create a new folder called redis-email-tutorial, then navigate to the folder created in your terminal. Copy, paste and run the following.

yarn init -y && yarn add dotenv express nodemailer @bull-board/api @bull-board/express

yarn add -D typescript nodemon eslint @typescript-eslint/parser @types/nodemailer @typescript-eslint/eslint-plugin @types/node @types/express ts-node

Create a new src folder then create an app.ts file in the folder created.
Copy and paste the following;

// src/app.ts

import express, { Express, Request, Response } from "express";
import dotenv from "dotenv";

dotenv.config();

const PORT = process.env.PORT || 3000;

const app: Express = express();

app.use(express.json());

app.post("/", async (req: Request, res: Response) => {
  res.send("Redis Email");
});

app.listen(PORT, () => {
  console.log(`\n Server is running at http://localhost:${PORT} ✅`);
});

Then edit the package.json file to contain the following scripts commands

  "scripts": {
    "dev": "nodemon --quiet src/app.ts",
    "start": "node build/src/index.js",
    "build": "tsc --project ./",
    "lint": "eslint src/**/*.ts",
    "format": "eslint src/**/*.ts --fix"
  },

You can now start the app in your terminal by running yarn dev in the root directory.

Setup Redis Server

To continue this process, please make sure you have docker installed on your machine. You can check this link Install Docker desktop on how to go about it.

We are now going to create a docker-compose.yml file and copy the following into it

# docker-compose.yml

version: "3.7"

services:
  redis:
    image: redis
    container_name: tutorial_redis_email
    environment:
      - ALLOW_EMPTY_PASSWORD=yes
    ports:
      - 6379:6379

  redis-commander:
    container_name: email-api-redis
    image: rediscommander/redis-commander
    restart: always
    environment:
      - REDIS_HOSTS=local:redis:6379
    ports:
      - "8080:8080"

Let's create a .env file to set the PORT and a REDIS_URL variable

# .env

PORT=3000
REDIS_URL=http://localhost:6379

We would ensure docker is completely setup on our local machine. Then we will open another instance of the terminal to start the docker container by running the following;

docker-compose up

There are other methods to get the redis server up and running, but to keep things simple and straighforward for both linux and windows developers, we would be going with the docker method.

Setting up Queues and Processes

Bull is a Node library that implements a fast and robust queue system based on redis. So let's quickly go through some basic terms

• Queue: A queue is a First-In-First-Out (FIFO) data structure and a queue instance can normally have 3 main different roles: A job producer, a job consumer or/and an events listener.
• Job Producer: A job producer is simply some Node program that adds jobs to a queue.
• Job Consumers: A job consumer or worker, is nothing more than a Node program that defines a process function. The queue executes every job with the consumer process.
• Event Listeners: This helps us to listen to events that happen in the queue. The Event listeners are attached to every process.

Let's illustrate this concept with a simple example - consider a Fuel Station equipped with a Fuel pump for drivers. In this scenario, we can identify the Fuel station as the producer since it is the destination for fuel whenever drivers require it, and it manages the queue of people waiting for their turn.

The Fuel pump, on the other hand, serves as our consumer or worker in this system. Its role is to dispense fuel to the drivers, treating the drivers' need for fuel as a Job with associated details such as the required quantity of fuel (JobData).

Moving forward, we can envision our event emitters as trackers responsible for monitoring whether drivers successfully receive the fuel they need. These trackers keep tabs on the entire process and can indicate outcomes like Success or Failure.

It is essential to note that fuel dispensing can occur anytime it's the car owner's turn, and during this time, other tasks can still be performed. For example, while providing feedback to the user, email requests can be held in a queue until they can be addressed.

Now lets get into setting up the queue.

We will be working with the folder structure shown above.
Starting with defining the queue, let's create a new folder queues in the src folder, then we will also create an index.ts file. Copy and paste the following into the file that was created.

// src/queues/index.ts

import Bull from "bull";
import dotenv from "dotenv";

dotenv.config();

export const emailQueue = new Bull("email", {
  redis: process.env.REDIS_URL,
});

This will define a queue for us called email and we will be using an instance of the queue to define our producers and consumers.
Then we will create a folder emails inside the queues directory to define our producers and consumers.
Create two files consumer.ts and producer.ts to continue below;

// src/queues/emails/producer.ts

import { emailQueue } from "..";
import { EmailData } from "./types";

export const sendMail = (data: EmailData) => {
  emailQueue.add(data, { attempts: 3 });
};

We have defined and exported a producer function to add our job to the email queue, now let's define our process to execute the job.

// src/queues/emails/consumer.ts

import { Job } from "bull";
import { EmailData } from "./types";
import nodemailer from "nodemailer";

const emailProcessor = async ({ data }: Job<EmailData>) => {
  try {
    const testMailAccount = await nodemailer.createTestAccount();

    const transporter = nodemailer.createTransport({
      host: "smtp.ethereal.email",
      port: 587,
      secure: false,
      auth: {
        user: testMailAccount.user, // generated ethereal user
        pass: testMailAccount.pass, // generated ethereal password
      },
      tls: {
        rejectUnauthorized: false,
      },
    });

    const response = await transporter.sendMail(data);

    console.log("Message sent to", data?.to);

    const responseUrl = nodemailer.getTestMessageUrl(response);

    console.log(responseUrl);

    return responseUrl;
  } catch (error) {
    throw new Error(error?.toString());
  }
};

export default emailProcessor;

This defines a function that sends a test email to an email address using nodemailer.

// src/queues/emails/types.ts

export type EmailData = {
  from: string;
  to: string;
  subject: string;
  html: string;
};

We have just defined the type of our job data, so lets proceed to attaching our queue to its consumer and defining some event listeners

// src/queues/process.ts

import { Job } from "bull";
import { emailQueue } from ".";
import emailProcessor from "./emails/consumer";

const handleFailure = (job: Job, err: Error) => {
  if (job.attemptsMade >= (job?.opts?.attempts || 0)) {
    console.log(
      `Job failures above threshold in ${job.queue.name} for: ${JSON.stringify(
        job.data
      )}`,
      err
    );
    job.remove();
    return null;
  }
  console.log(
    `Job in ${job.queue.name} failed for: ${JSON.stringify(job.data)} with ${
      err.message
    }. ${(job?.opts?.attempts || 0) - job.attemptsMade} attempts left`
  );
};

const handleCompleted = (job: Job) => {
  console.log(
    `Job in ${job.queue.name} completed for: ${JSON.stringify(job.data)}`
  );
  job.remove();
};

const handleStalled = (job: Job) => {
  console.log(
    `Job in ${job.queue.name} stalled for: ${JSON.stringify(job.data)}`
  );
};

export const startQueues = () => {
  try {
    emailQueue.process(emailProcessor);
    emailQueue.on("failed", handleFailure);
    emailQueue.on("completed", handleCompleted);
    emailQueue.on("stalled", handleStalled);
    console.log("\n Queue jobs started successfully ✅");
  } catch (err) {
    console.log("Queue Jobs Error: ", err);
  }
};

We have attached our job consumer to our queue, we can now define an adapter so we can view our queue processes. We will be using bull-board which is one of some other third-party UIs that you can use for monitoring your queues. Check here for the Bull Package

// src/queues/adapter.ts
import { createBullBoard } from "@bull-board/api";
import { BullAdapter } from "@bull-board/api/bullAdapter";
import { ExpressAdapter } from "@bull-board/express";
import { emailQueue } from ".";

const serverAdapter = new ExpressAdapter();

createBullBoard({
  queues: [new BullAdapter(emailQueue)],
  serverAdapter,
});

export default serverAdapter;

Finally lets start our queues from our app.ts file, edit your app.ts file to contain the following

// src/app.ts

import express, { Express, Request, Response } from "express";
import dotenv from "dotenv";
import { sendMail } from "./queues/emails/producer";
import { startQueues } from "./queues/process";
import serverAdapter from "./queues/adapter";

dotenv.config();

const app: Express = express();
const PORT = process.env.PORT;

app.use(express.json());

startQueues();

app.post("/mail", async (req: Request, res: Response) => {
  const { message, ...emailData } = req.body;
  await sendMail({ ...emailData, html: `<p>${message}</p>` });
  res.send("Sent Email");
});

app.use("/admin/queues", serverAdapter.getRouter());

app.listen(PORT, () => {
  console.log(`\n Server is running at http://localhost:${PORT} ✅`);
});

So we imported our startQueues function to start our queues and we edited the /mail endpoint to pass a job object into our email producer.
We also added a new path /admin/queues for us to inspect our queues.
If you go to the endpoint you should see your job data in your console.

Connecting to a free redis database

To setup a redis database for production, you can check the available services provided by your cloud server e.g Heroku Redis, Redis.com etc. We would be using the redis.com free database for this tutorial.
Visit redis.com, create an account and proceed to create a free redis database. You should have something like this;

Click on connect, then we will copy the url displayed in the RedisInsight connection box.

Click on copy to get the full url or replace the url should with the username and password for your account. It will be in this format.

# sample link

redis://default:[e-mail]:18364

After we have copied the link, we will replace our REDIS_URL environment variable to the link copied

# .env 

PORT=3000
REDIS_URL=redis://<username>:<password>@redis-13975.c81.us-east-1-2.ec2.cloud.redislabs.com:13975

Then you can stop your docker instance and restart your app, any request send to the /mail endpoint should still work correctly with the remote redis server.

Conclusion

We have been able setup a redis queue to send our emails. Queues can be used for a wide range of functions, you can take your time to study these queues and how they can help you to seperate some jobs on your server.

You can take a look at the Github Repository

References

• https://levelup.gitconnected.com/how-to-implement-queues-in-node-js-8b3a06ce0dd0
• https://github.com/codingvenue/nodejs-queue-redis-bull

Comments

[OYERINDE DANIEL]

Nice article.