AWS Messaging

Amazon SQS

Simple Queue Service provides managed message queues. Standard queues offer unlimited throughput with at-least-once delivery. FIFO queues guarantee exactly-once processing and ordering.

Producers send messages; consumers poll and process. Visibility timeout hides messages during processing. Dead letter queues (DLQ) capture messages that fail processing repeatedly.

• Standard queue: best-effort ordering, at-least-once delivery
• FIFO queue: strict ordering, exactly-once, 300 msg/sec limit
• Long polling (WaitTimeSeconds: 20) reduces empty responses and cost

// Send message to SQS
import { SQSClient, SendMessageCommand } from '@aws-sdk/client-sqs';

const sqs = new SQSClient({ region: 'us-east-1' });
await sqs.send(new SendMessageCommand({
  QueueUrl: 'https://sqs.us-east-1.amazonaws.com/123/orders',
  MessageBody: JSON.stringify({ orderId: '123', action: 'process' }),
}));

Amazon SNS

Simple Notification Service provides pub/sub messaging. Publishers send to topics; subscribers receive via SQS, Lambda, HTTP, email, or SMS. Fan-out pattern: one message reaches multiple consumers.

Combine SNS + SQS for reliable fan-out — SNS delivers to SQS queues, each consumed independently. Message filtering allows subscribers to receive only relevant messages.

# Create SNS topic with SQS subscription
aws sns create-topic --name order-events
aws sns subscribe \
  --topic-arn arn:aws:sns:us-east-1:123:order-events \
  --protocol sqs \
  --notification-endpoint arn:aws:sqs:us-east-1:123:email-worker

# Filter policy: only order.completed events
aws sns set-subscription-attributes \
  --subscription-arn arn:aws:sns:... \
  --attribute-name FilterPolicy \
  --attribute-value '{"eventType": ["order.completed"]}'

Amazon EventBridge

EventBridge is a serverless event bus connecting AWS services, SaaS applications, and custom applications. Rules match event patterns and route to targets (Lambda, SQS, Step Functions, API destinations).

Use EventBridge for event-driven architectures: decouple services, react to AWS API calls (via CloudTrail), schedule recurring tasks, and integrate third-party SaaS events.

# EventBridge rule: S3 upload triggers processing
{
  "source": ["aws.s3"],
  "detail-type": ["Object Created"],
  "detail": {
    "bucket": {"name": ["uploads"]},
    "object": {"key": [{"suffix": ".csv"}]}
  }
}

# Schedule: run every day at 2 AM UTC
# ScheduleExpression: cron(0 2 * * ? *)

Amazon Kinesis

Kinesis Data Streams ingests real-time data at scale — clickstreams, IoT sensor data, application logs. Kinesis Data Firehose delivers streams to S3, Redshift, OpenSearch, and third-party services.

Data Streams: custom consumers with shard-based parallelism. Firehose: fully managed delivery with optional transformation. Choose Streams for custom processing, Firehose for delivery pipelines.

# Put record to Kinesis stream
aws kinesis put-record \
  --stream-name clickstream \
  --partition-key user-123 \
  --data "$(echo '{"page": "/home", "timestamp": "2024-01-15T10:00:00Z"}' | base64)"

Messaging Architecture Patterns

Choose the right service: SQS for task queues and work distribution. SNS for notifications and fan-out. EventBridge for event routing and scheduling. Kinesis for real-time streaming analytics.

Implement idempotent consumers — messages may be delivered more than once. Use dead letter queues for failed messages. Monitor queue depth and age of oldest message as key metrics.

# Pattern selection
# Task queue (one consumer per message) → SQS
# Fan-out (multiple consumers) → SNS + SQS
# Event routing (pattern matching) → EventBridge
# Real-time streaming → Kinesis Data Streams
# Data delivery pipeline → Kinesis Firehose
# Cross-service async calls → SQS or EventBridge