sqs-consumer

Build SQS-based applications without the boilerplate. Just define an async function that handles the SQS message processing.

Context

from @SishaarRao Hello world! This is my Lob Halloween Hackathon Project. sqs-consumer is an awesome library used by many Node developers and Lob engineering. It provides a lot of boilerplate functionality in a simple yet powerful manner.

That being said, nowadays we @ Lob opt away from this package, primarily because it has not yet been ported to the new AWS SDK V3.

I forked the sqs-consumer in order to port the repo, and make some improvements we've wanted to see.

Next challenges

1. Investigate why sqs-consumer doesn't play nice with FIFO queues

Changelog (10-13-2022)

• Upgrade all dependencies to latest versions
• Upgrade underlying AWS SDK to V3
• Convert test suite to Jest

Installation

npm install @lob/sqs-consumer --save

Usage

const { Consumer } = require('sqs-consumer');

const app = Consumer.create({
  queueUrl: 'https://sqs.eu-west-1.amazonaws.com/account-id/queue-name',
  handleMessage: async (message) => {
    // do some work with `message`
  }
});

app.on('error', (err) => {
  console.error(err.message);
});

app.on('processing_error', (err) => {
  console.error(err.message);
});

app.start();

• The queue is polled continuously for messages using long polling.
• Messages are deleted from the queue once the handler function has completed successfully.
• Throwing an error (or returning a rejected promise) from the handler function will cause the message to be left on the queue. An SQS redrive policy can be used to move messages that cannot be processed to a dead letter queue.
• By default messages are processed one at a time – a new message won't be received until the first one has been processed. To process messages in parallel, use the batchSize option detailed below.
• By default, the default Node.js HTTP/HTTPS SQS agent creates a new TCP connection for every new request (AWS SQS documentation). To avoid the cost of establishing a new connection, you can reuse an existing connection by passing a new SQS instance with keepAlive: true.

const { Consumer } = require('sqs-consumer');
const AWS = require('aws-sdk');
const https = require('https');

const app = Consumer.create({
  queueUrl: 'https://sqs.eu-west-1.amazonaws.com/account-id/queue-name',
  handleMessage: async (message) => {
    // do some work with `message`
  },
  sqs: new AWS.SQS({
    httpOptions: {
      agent: new https.Agent({
        keepAlive: true
      })
    }
  })
});

app.on('error', (err) => {
  console.error(err.message);
});

app.on('processing_error', (err) => {
  console.error(err.message);
});

app.start();

Credentials

By default the consumer will look for AWS credentials in the places specified by the AWS SDK. The simplest option is to export your credentials as environment variables:

export AWS_SECRET_ACCESS_KEY=...
export AWS_ACCESS_KEY_ID=...

If you need to specify your credentials manually, you can use a pre-configured instance of the AWS SQS client:

const { Consumer } = require('sqs-consumer');
const AWS = require('aws-sdk');

AWS.config.update({
  region: 'eu-west-1',
  accessKeyId: '...',
  secretAccessKey: '...'
});

const app = Consumer.create({
  queueUrl: 'https://sqs.eu-west-1.amazonaws.com/account-id/queue-name',
  handleMessage: async (message) => {
    // ...
  },
  sqs: new AWS.SQS()
});

app.on('error', (err) => {
  console.error(err.message);
});

app.on('processing_error', (err) => {
  console.error(err.message);
});

app.on('timeout_error', (err) => {
 console.error(err.message);
});

app.start();

API

Consumer.create(options)

Creates a new SQS consumer.

Options

• queueUrl - String - The SQS queue URL
• region - String - The AWS region (default eu-west-1)
• handleMessage - Function - An async function (or function that returns a Promise) to be called whenever a message is received. Receives an SQS message object as it's first argument.
• handleMessageBatch - Function - An async function (or function that returns a Promise) to be called whenever a batch of messages is received. Similar to handleMessage but will receive the list of messages, not each message individually. If both are set, handleMessageBatch overrides handleMessage.
• handleMessageTimeout - Number - Time in ms to wait for handleMessage to process a message before timing out. Emits timeout_error on timeout. By default, if handleMessage times out, the unprocessed message returns to the end of the queue.
• attributeNames - Array - List of queue attributes to retrieve (i.e. ['All', 'ApproximateFirstReceiveTimestamp', 'ApproximateReceiveCount']).
• messageAttributeNames - Array - List of message attributes to retrieve (i.e. ['name', 'address']).
• batchSize - Number - The number of messages to request from SQS when polling (default 1). This cannot be higher than the AWS limit of 10.
• visibilityTimeout - Number - The duration (in seconds) that the received messages are hidden from subsequent retrieve requests after being retrieved by a ReceiveMessage request.
• heartbeatInterval - Number - The interval (in seconds) between requests to extend the message visibility timeout. On each heartbeat the visibility is extended by adding visibilityTimeout to the number of seconds since the start of the handler function. This value must less than visibilityTimeout.
• terminateVisibilityTimeout - Boolean - If true, sets the message visibility timeout to 0 after a processing_error (defaults to false).
• waitTimeSeconds - Number - The duration (in seconds) for which the call will wait for a message to arrive in the queue before returning.
• authenticationErrorTimeout - Number - The duration (in milliseconds) to wait before retrying after an authentication error (defaults to 10000).
• pollingWaitTimeMs - Number - The duration (in milliseconds) to wait before repolling the queue (defaults to 0).
• sqs - Object - An optional AWS SQS object to use if you need to configure the client manually
• shouldDeleteMessages - Boolean - Default to true, if you don't want the package to delete messages from sqs set this to false

consumer.start()

Start polling the queue for messages.

consumer.stop()

Stop polling the queue for messages.

consumer.isRunning

Returns the current polling state of the consumer: true if it is actively polling, false if it is not.

Events

Each consumer is an EventEmitter and emits the following events:

|Event|Params|Description| |-----|------|-----------| |error|err, [message]|Fired when an error occurs interacting with the queue. If the error correlates to a message, that message is included in Params| |processing_error|err, message|Fired when an error occurs processing the message.| |timeout_error|err, message|Fired when handleMessageTimeout is supplied as an option and if handleMessage times out.| |message_received|message|Fired when a message is received.| |message_processed|message|Fired when a message is successfully processed and removed from the queue.| |response_processed|None|Fired after one batch of items (up to batchSize) has been successfully processed.| |stopped|None|Fired when the consumer finally stops its work.| |empty|None|Fired when the queue is empty (All messages have been consumed).|

AWS IAM Permissions

Consumer will receive and delete messages from the SQS queue. Ensure sqs:ReceiveMessage, sqs:DeleteMessage, sqs:DeleteMessageBatch, sqs:ChangeMessageVisibility and sqs:ChangeMessageVisibilityBatch access is granted on the queue being consumed.

Contributing

See contributing guidelines.