AGENT SKILL

message-queues

Message queue best practices and patterns

License 许可
MIT
最近更新
2026-03-01
Stars
★ 3
Forks
2

01 前言 Front Matter

name
message-queues
license
MIT
metadata
{ "audience": "developers", "category": "architecture" }
description
Message queue best practices and patterns
compatibility
opencode

03 SKILL.md 正文

What I do

  • Implement reliable message publishing and consuming
  • Handle message ordering and deduplication
  • Use dead letter queues for failed messages
  • Implement retry with exponential backoff
  • Use topic exchanges for pub/sub patterns
  • Handle consumer groups and load balancing
  • Design for at-least-once or exactly-once semantics
  • Monitor queue health and lag

When to use me

When implementing message queue systems (RabbitMQ, Kafka, SQS).

RabbitMQ Patterns

import pika
import json
from typing import Callable
from dataclasses import dataclass
from datetime import datetime


@dataclass
class Message:
    id: str
    type: str
    payload: dict
    timestamp: datetime
    retry_count: int = 0


class RabbitMQPublisher:
    def __init__(
        self,
        host: str = "localhost",
        port: int = 5672,
        username: str = "guest",
        password: str = "guest",
    ) -> None:
        self.host = host
        self.port = port
        self.credentials = pika.PlainCredentials(username, password)
        self.connection = None
        self.channel = None

    def connect(self):
        self.connection = pika.BlockingConnection(
            pika.ConnectionParameters(
                host=self.host,
                port=self.port,
                credentials=self.credentials,
                heartbeat=600,
                blocked_connection_timeout=300,
            )
        )
        self.channel = self.connection.channel()

        # Declare exchanges
        self.channel.exchange_declare(
            exchange='events',
            exchange_type='topic',
            durable=True
        )

        # Declare dead letter exchange
        self.channel.exchange_declare(
            exchange='events.dlx',
            exchange_type='direct',
            durable=True
        )

    def publish(
        self,
        routing_key: str,
        message: Message,
        exchange: str = 'events'
    ) -> None:
        if not self.connection or self.connection.is_closed:
            self.connect()

        properties = pika.BasicProperties(
            delivery_mode=2,  # Persistent
            content_type='application/json',
            message_id=message.id,
            timestamp=int(message.timestamp.timestamp()),
            headers={
                'retry_count': message.retry_count,
                'x-dead-letter-exchange': 'events.dlx',
                'x-dead-letter-routing-key': f'{routing_key}.dlq',
            }
        )

        self.channel.basic_publish(
            exchange=exchange,
            routing_key=routing_key,
            body=json.dumps(message.__dict__),
            properties=properties,
        )

    def close(self):
        if self.connection and not self.connection.is_closed:
            self.connection.close()


class RabbitMQConsumer:
    def __init__(
        self,
        queue_name: str,
        handler: Callable[[Message], bool],
        prefetch_count: int = 10,
    ) -> None:
        self.queue_name = queue_name
        self.handler = handler
        self.prefetch_count = prefetch_count

    def start(self):
        connection = pika.BlockingConnection(
            pika.ConnectionParameters(host='localhost')
        )
        channel = connection.channel()

        # Declare queue with dead letter config
        channel.queue_declare(
            queue=self.queue_name,
            durable=True,
            arguments={
                'x-dead-letter-exchange': 'events.dlx',
                'x-dead-letter-routing-key': f'{self.queue_name}.dlq',
                'x-message-ttl': 86400000,  # 24 hours
            }
        )

        # Dead letter queue
        channel.queue_declare(
            queue=f'{self.queue_name}.dlq',
            durable=True,
        )

        channel.basic_qos(prefetch_count=self.prefetch_count)
        channel.basic_consume(
            queue=self.queue_name,
            on_message_callback=self._handle_message,
        )

        print(f'Consumer started on queue: {self.queue_name}')
        channel.start_consuming()

    def _handle_message(self, channel, method, properties, body):
        try:
            message = Message(**json.loads(body))

            success = self.handler(message)

            if success:
                channel.basic_ack(delivery_tag=method.delivery_tag)
            else:
                # Requeue for retry
                channel.basic_nack(
                    delivery_tag=method.delivery_tag,
                    requeue=False,
                )

        except Exception as e:
            print(f"Error processing message: {e}")
            channel.basic_nack(
                delivery_tag=method.delivery_tag,
                requeue=False,
            )

Kafka Patterns

from kafka import KafkaProducer, KafkaConsumer
from kafka.errors import KafkaError
import json


class KafkaPublisher:
    def __init__(self, bootstrap_servers: list) -> None:
        self.producer = KafkaProducer(
            bootstrap_servers=bootstrap_servers,
            value_serializer=lambda v: json.dumps(v).encode('utf-8'),
            key_serializer=lambda k: k.encode('utf-8') if k else None,
            acks='all',  # Wait for all replicas
            retries=3,
            retry_backoff_ms=500,
            linger_ms=10,  # Batch messages
            batch_size=16384,
        )

    def publish(
        self,
        topic: str,
        key: str,
        value: dict,
        partition_key: str = None
    ) -> None:
        try:
            future = self.producer.send(
                topic,
                key=key,
                value=value,
                timestamp_ms=int(datetime.utcnow().timestamp() * 1000),
            )

            # Wait for confirmation
            record_metadata = future.get(timeout=10)
            print(f"Message sent to {record_metadata.topic}[{record_metadata.partition}]")

        except KafkaError as e:
            print(f"Failed to send message: {e}")
            raise


class KafkaConsumer:
    def __init__(
        self,
        bootstrap_servers: list,
        group_id: str,
        topics: list,
        auto_offset_reset: str = 'earliest',
    ) -> None:
        self.consumer = KafkaConsumer(
            *topics,
            bootstrap_servers=bootstrap_servers,
            group_id=group_id,
            auto_offset_reset=auto_offset_reset,
            enable_auto_commit=False,
            value_deserializer=lambda v: json.loads(v.decode('utf-8')),
            key_deserializer=lambda k: k.decode('utf-8') if k else None,
        )

    def consume(self, handler: callable):
        for message in self.consumer:
            try:
                handler(
                    topic=message.topic,
                    partition=message.partition,
                    offset=message.offset,
                    key=message.key,
                    value=message.value,
                )

                # Manual commit after successful processing
                self.consumer.commit()

            except Exception as e:
                print(f"Error processing message: {e}")
                # Implement retry logic or dead letter

Retry with Exponential Backoff

import time
from functools import wraps


def retry_with_backoff(
    max_retries: int = 3,
    initial_delay: float = 1.0,
    max_delay: float = 60.0,
    exponential_base: float = 2.0,
):
    def decorator(func):
        @wraps(func)
        def wrapper(*args, **kwargs):
            retries = 0
            delay = initial_delay

            while True:
                try:
                    return func(*args, **kwargs)
                except Exception as e:
                    if retries >= max_retries:
                        raise

                    retries += 1

                    time.sleep(delay)
                    delay = min(delay * exponential_base, max_delay)

                    print(f"Retry {retries}/{max_retries} after {delay:.2f}s")

        return wrapper
    return decorator

SQS with Lambda

import boto3
import json
from datetime import datetime


sqs = boto3.client('sqs')
queue_url = 'https://sqs.us-east-1.amazonaws.com/123456789012/my-queue'


def send_message(message: dict, delay_seconds: int = 0) -> str:
    response = sqs.send_message(
        QueueUrl=queue_url,
        MessageBody=json.dumps(message),
        DelaySeconds=delay_seconds,
        MessageAttributes={
            'timestamp': {
                'StringValue': datetime.utcnow().isoformat(),
                'DataType': 'String',
            }
        }
    )
    return response['MessageId']


def receive_messages(max_messages: int = 10) -> list:
    response = sqs.receive_message(
        QueueUrl=queue_url,
        MaxNumberOfMessages=max_messages,
        WaitTimeSeconds=20,
        MessageAttributeNames=['All'],
    )
    return response.get('Messages', [])


def delete_message(receipt_handle: str):
    sqs.delete_message(
        QueueUrl=queue_url,
        ReceiptHandle=receipt_handle,
    )

04 同仓库更多 Skills