AWS SQS Consumers

Q: sqs-consumer npm package?

sqs-consumer wraps polling with events. Fine for ECS workers. Understand visibility semantics either way.

Q: How many pollers?

Scale consumers on ApproximateNumberOfMessagesVisible metric. Avoid duplicate tight loops without long poll.

Q: IAM for worker?

sqs:ReceiveMessage, DeleteMessage, ChangeMessageVisibility on queue ARN only.

Q: Express API enqueue?

SendMessageCommand from API with small JSON body. Same idempotency rules as BullMQ producer.

Amazon SQS gives managed queues without operating Redis. Node workers use AWS SDK v3 with long polling, visibility timeouts, and dead-letter queues (DLQ).

Recipe

Quick-reference recipe card - copy-paste ready.

import {
  SQSClient,
  ReceiveMessageCommand,
  DeleteMessageCommand,
} from "@aws-sdk/client-sqs";
 
const sqs = new SQSClient({});
const queueUrl = process.env.SQS_QUEUE_URL!;
 
const res = await sqs.send(
  new ReceiveMessageCommand({
    QueueUrl: queueUrl,
    MaxNumberOfMessages: 10,
    WaitTimeSeconds: 20,
    VisibilityTimeout: 60,
  })
);
 
for (const msg of res.Messages ?? []) {
  await processBody(msg.Body!);
  await sqs.send(
    new DeleteMessageCommand({ QueueUrl: queueUrl, ReceiptHandle: msg.ReceiptHandle! })
  );
}

When to reach for this:

AWS-native stack without Redis
Need managed scaling and pay-per-message pricing
Cross-account or cross-region fan-out with SNS

Working Example

// src/sqs/poll.ts
import {
  SQSClient,
  ReceiveMessageCommand,
  DeleteMessageCommand,
  ChangeMessageVisibilityCommand,
} from "@aws-sdk/client-sqs";
 
const sqs = new SQSClient({ region: process.env.AWS_REGION });
const queueUrl = process.env.SQS_QUEUE_URL!;
 
async function handleMessage(body: string) {
  const payload = JSON.parse(body) as { type: string; orderId: string };
  if (payload.type === "fulfill") {
    await fulfillOrder(payload.orderId);
  }
}
 
export async function pollForever(signal: AbortSignal) {
  while (!signal.aborted) {
    const res = await sqs.send(
      new ReceiveMessageCommand({
        QueueUrl: queueUrl,
        MaxNumberOfMessages: 10,
        WaitTimeSeconds: 20,
        VisibilityTimeout: 120,
        MessageAttributeNames: ["All"],
      })
    );
 
    for (const msg of res.Messages ?? []) {
      try {
        await handleMessage(msg.Body ?? "{}");
        await sqs.send(
          new DeleteMessageCommand({
            QueueUrl: queueUrl,
            ReceiptHandle: msg.ReceiptHandle!,
          })
        );
      } catch (err) {
        console.error({ msg: "sqs_handler_error", messageId: msg.MessageId, err });
        // Message returns to queue after visibility timeout
      }
    }
  }
}
 
// src/worker-main.ts
const ac = new AbortController();
process.on("SIGTERM", () => ac.abort());
await pollForever(ac.signal);

What this demonstrates:

Long poll WaitTimeSeconds: 20 reduces cost and CPU spin
Delete only after successful handler
Failed handler relies on visibility timeout retry (plus DLQ at infra level)

Deep Dive

Visibility Timeout Math

visibility_timeout >= p99_processing_time * 1.5

Too short: duplicate processing while first handler still runs
Too long: slow retry on failure
Extend with ChangeMessageVisibility for variable-length jobs

Dead-Letter Queue (Infrastructure)

{
  "RedrivePolicy": {
    "deadLetterTargetArn": "arn:aws:sqs:...:dlq",
    "maxReceiveCount": 5
  }
}

After 5 failed receives, message moves to DLQ
Runbook: inspect DLQ, fix bug, redrive messages

FIFO vs Standard

Type	Ordering	Throughput	Dedup
Standard	Best-effort	Very high	App-level idempotency
FIFO	Per message group	300 TPS/group	Optional content dedup

Payments and inventory often use FIFO with MessageGroupId

// Message body may be SNS wrapper JSON - unwrap Envelope
const outer = JSON.parse(body);
const inner = outer.Message ? JSON.parse(outer.Message) : outer;

SDK v3 Batch Partial Failure (Lambda)

Lambda SQS event source mapping supports batch item failures
ECS/EC2 long poll loop handles one message at a time or manual batch ack

Gotchas

Delete before work completes - message loss. Fix: delete last.
Visibility shorter than job - duplicate fulfillments. Fix: extend visibility or raise timeout.
No DLQ - poison message infinite loop. Fix: maxReceiveCount + DLQ alarm.
Parsing SNS wrapper wrong - silent no-op handlers. Fix: unwrap Message field.
Ignoring idempotency - at-least-once delivery duplicates work. Fix: idempotency keys in DB.

Alternatives

Alternative	Use When	Don't Use When
BullMQ	Redis already running, rich job API	Want zero queue infra
Kinesis	Streaming analytics	Simple task queue
EventBridge	Event routing rules	Point-to-point worker queue only
Google Pub/Sub	GCP stack	AWS-only org

FAQs

sqs-consumer npm package?

sqs-consumer wraps polling with events. Fine for ECS workers. Understand visibility semantics either way.

How many pollers?

Scale consumers on ApproximateNumberOfMessagesVisible metric. Avoid duplicate tight loops without long poll.

Exactly-once?

No. Standard queue is at-least-once. Design idempotent handlers.

Large payloads?

Use S3 extended client pattern - SQS carries S3 pointer when >256KB.

IAM for worker?

sqs:ReceiveMessage, DeleteMessage, ChangeMessageVisibility on queue ARN only.

Local dev?

LocalStack or ElasticMQ for integration tests. ElasticMQ docker for laptop.

Express API enqueue?

SendMessageCommand from API with small JSON body. Same idempotency rules as BullMQ producer.

Message retention?

Default 4 days. Raise to 14 for replay windows during outages.

FIFO throughput?

Batching and multiple message group ids scale parallel ordered streams.

Cost drivers?

Requests count. Long polling reduces empty receives. Batch receive up to 10 messages.

Queues Basics - async mental model
AWS SDK v3 - client setup
Idempotency Keys - duplicate delivery
BullMQ - Redis alternative
Queues Best Practices - DLQ runbooks

Node.js Best Practices (GitHub) - Curated production best practices.
DigitalOcean Node.js Tutorials - Production-oriented how-tos.
NodeSource Blog & Resources - Enterprise Node.js performance and security content.
Distributed Systems with Node.js - Enterprise-ready distributed systems.
Node.js Design Patterns, 4th Edition - Production-grade application patterns.

Stack versions: This page was written for Node.js 24.18.0 (Active LTS), npm 10+, TypeScript 5.6+, Express 5, Fastify 5, and NestJS 11.

Recipe

Quick-reference recipe card - copy-paste ready.

import {
  SQSClient,
  ReceiveMessageCommand,
  DeleteMessageCommand,
} from "@aws-sdk/client-sqs";
 
const sqs = new SQSClient({});
const queueUrl = process.env.SQS_QUEUE_URL!;
 
const res = await sqs.send(
  new ReceiveMessageCommand({
    QueueUrl: queueUrl,
    MaxNumberOfMessages: 10,
    WaitTimeSeconds: 20,
    VisibilityTimeout: 60,
  })
);
 
for (const msg of res.Messages ?? []) {
  await processBody(msg.Body!);
  await sqs.send(
    new DeleteMessageCommand({ QueueUrl: queueUrl, ReceiptHandle: msg.ReceiptHandle! })
  );
}

When to reach for this:

AWS-native stack without Redis
Need managed scaling and pay-per-message pricing
Cross-account or cross-region fan-out with SNS

Working Example

// src/sqs/poll.ts
import {
  SQSClient,
  ReceiveMessageCommand,
  DeleteMessageCommand,
  ChangeMessageVisibilityCommand,
} from "@aws-sdk/client-sqs";
 
const sqs = new SQSClient({ region: process.env.AWS_REGION });
const queueUrl = process.env.SQS_QUEUE_URL!;
 
async function handleMessage(body: string) {
  const payload = JSON.parse(body) as { type: string; orderId: string };
  if (payload.type === "fulfill") {
    await fulfillOrder(payload.orderId);
  }
}
 
export async function pollForever(signal: AbortSignal) {
  while (!signal.aborted) {
    const res = await sqs.send(
      new ReceiveMessageCommand({
        QueueUrl: queueUrl,
        MaxNumberOfMessages: 10,
        WaitTimeSeconds: 20,
        VisibilityTimeout: 120,
        MessageAttributeNames: ["All"],
      })
    );
 
    for (const msg of res.Messages ?? []) {
      try {
        await handleMessage(msg.Body ?? "{}");
        await sqs.send(
          new DeleteMessageCommand({
            QueueUrl: queueUrl,
            ReceiptHandle: msg.ReceiptHandle!,
          })
        );
      } catch (err) {
        console.error({ msg: "sqs_handler_error", messageId: msg.MessageId, err });
        // Message returns to queue after visibility timeout
      }
    }
  }
}
 
// src/worker-main.ts
const ac = new AbortController();
process.on("SIGTERM", () => ac.abort());
await pollForever(ac.signal);

What this demonstrates:

Long poll WaitTimeSeconds: 20 reduces cost and CPU spin
Delete only after successful handler
Failed handler relies on visibility timeout retry (plus DLQ at infra level)

Deep Dive

Visibility Timeout Math

visibility_timeout >= p99_processing_time * 1.5

Too short: duplicate processing while first handler still runs
Too long: slow retry on failure
Extend with ChangeMessageVisibility for variable-length jobs

Dead-Letter Queue (Infrastructure)

{
  "RedrivePolicy": {
    "deadLetterTargetArn": "arn:aws:sqs:...:dlq",
    "maxReceiveCount": 5
  }
}

After 5 failed receives, message moves to DLQ
Runbook: inspect DLQ, fix bug, redrive messages

FIFO vs Standard

Type	Ordering	Throughput	Dedup
Standard	Best-effort	Very high	App-level idempotency
FIFO	Per message group	300 TPS/group	Optional content dedup

Payments and inventory often use FIFO with MessageGroupId

// Message body may be SNS wrapper JSON - unwrap Envelope
const outer = JSON.parse(body);
const inner = outer.Message ? JSON.parse(outer.Message) : outer;

SDK v3 Batch Partial Failure (Lambda)

Lambda SQS event source mapping supports batch item failures
ECS/EC2 long poll loop handles one message at a time or manual batch ack

Gotchas

Delete before work completes - message loss. Fix: delete last.
Visibility shorter than job - duplicate fulfillments. Fix: extend visibility or raise timeout.
No DLQ - poison message infinite loop. Fix: maxReceiveCount + DLQ alarm.
Parsing SNS wrapper wrong - silent no-op handlers. Fix: unwrap Message field.
Ignoring idempotency - at-least-once delivery duplicates work. Fix: idempotency keys in DB.

Alternatives

Alternative	Use When	Don't Use When
BullMQ	Redis already running, rich job API	Want zero queue infra
Kinesis	Streaming analytics	Simple task queue
EventBridge	Event routing rules	Point-to-point worker queue only
Google Pub/Sub	GCP stack	AWS-only org

FAQs

sqs-consumer npm package?

sqs-consumer wraps polling with events. Fine for ECS workers. Understand visibility semantics either way.

How many pollers?

Scale consumers on ApproximateNumberOfMessagesVisible metric. Avoid duplicate tight loops without long poll.

Exactly-once?

No. Standard queue is at-least-once. Design idempotent handlers.

Large payloads?

Use S3 extended client pattern - SQS carries S3 pointer when >256KB.

IAM for worker?

sqs:ReceiveMessage, DeleteMessage, ChangeMessageVisibility on queue ARN only.

Local dev?

LocalStack or ElasticMQ for integration tests. ElasticMQ docker for laptop.

Express API enqueue?

SendMessageCommand from API with small JSON body. Same idempotency rules as BullMQ producer.

Message retention?

Default 4 days. Raise to 14 for replay windows during outages.

FIFO throughput?

Batching and multiple message group ids scale parallel ordered streams.

Cost drivers?

Requests count. Long polling reduces empty receives. Batch receive up to 10 messages.

Queues Basics - async mental model
AWS SDK v3 - client setup
Idempotency Keys - duplicate delivery
BullMQ - Redis alternative
Queues Best Practices - DLQ runbooks

Node.js Best Practices (GitHub) - Curated production best practices.
DigitalOcean Node.js Tutorials - Production-oriented how-tos.
NodeSource Blog & Resources - Enterprise Node.js performance and security content.
Distributed Systems with Node.js - Enterprise-ready distributed systems.
Node.js Design Patterns, 4th Edition - Production-grade application patterns.

Stack versions: This page was written for Node.js 24.18.0 (Active LTS), npm 10+, TypeScript 5.6+, Express 5, Fastify 5, and NestJS 11.

AWS SQS Consumers

Search Documentation

AWS SQS Consumers