Esta página foi traduzida pela API Cloud Translation.

Como receber mensagens usando pull

O Pub/Sub é compatível com entrega de mensagens por push e pull. Para ter uma visão geral e comparar assinaturas de pull e push, consulte esta página. Este documento descreve a entrega por pull. Para ver uma discussão da entrega por push, consulte o Guia do assinante de push.

Pull assíncrono

O uso do pull assíncrono oferece uma maior capacidade em seu aplicativo, não exigindo que ele bloqueie novas mensagens. As mensagens podem ser recebidas no aplicativo usando-se um detector de mensagens de longa duração e reconhecidas uma de cada vez, conforme mostrado no exemplo abaixo. Os clientes Java, Python, .NET, Go e Ruby usam a API de serviço streamingPull para implementar a API do cliente assíncrona com eficiência.

Nem todas as bibliotecas de cliente são compatíveis com mensagens de pull assíncrono. Para saber mais sobre mensagens de pull síncrono, consulte Pull síncrono.

Para ver mais informações, consulte a documentação de referência da API da sua linguagem de programação.

C++

Antes de tentar esse exemplo, siga as instruções de configuração do C++ em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub C++.

Veja no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
auto sample = [](pubsub::Subscriber subscriber) {
  return subscriber.Subscribe(
      [&](pubsub::Message const& m, pubsub::AckHandler h) {
        std::cout << "Received message " << m << "\n";
        std::move(h).ack();
        PleaseIgnoreThisSimplifiesTestingTheSamples();
      });
};

C#

Antes de tentar esse exemplo, siga as instruções de configuração do C# em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub C#.

Veja no GitHub Feedback


using Google.Cloud.PubSub.V1;
using System;
using System.Linq;
using System.Text;
using System.Threading;
using System.Threading.Tasks;

public class PullMessagesAsyncSample
{
    public async Task<int> PullMessagesAsync(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);
        SubscriberClient subscriber = await SubscriberClient.CreateAsync(subscriptionName);
        // SubscriberClient runs your message handle function on multiple
        // threads to maximize throughput.
        int messageCount = 0;
        Task startTask = subscriber.StartAsync((PubsubMessage message, CancellationToken cancel) =>
        {
            string text = System.Text.Encoding.UTF8.GetString(message.Data.ToArray());
            Console.WriteLine($"Message {message.MessageId}: {text}");
            Interlocked.Increment(ref messageCount);
            return Task.FromResult(acknowledge ? SubscriberClient.Reply.Ack : SubscriberClient.Reply.Nack);
        });
        // Run for 5 seconds.
        await Task.Delay(5000);
        await subscriber.StopAsync(CancellationToken.None);
        // Lets make sure that the start task finished successfully after the call to stop.
        await startTask;
        return messageCount;
    }
}

Go

Antes de tentar esse exemplo, siga as instruções de configuração do Go em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub Go.

Veja no GitHub Feedback

import (
	"context"
	"fmt"
	"io"
	"sync"

	"cloud.google.com/go/pubsub"
)

func pullMsgs(w io.Writer, projectID, subID string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}
	defer client.Close()

	// Consume 10 messages.
	var mu sync.Mutex
	received := 0
	sub := client.Subscription(subID)
	cctx, cancel := context.WithCancel(ctx)
	err = sub.Receive(cctx, func(ctx context.Context, msg *pubsub.Message) {
		mu.Lock()
		defer mu.Unlock()
		fmt.Fprintf(w, "Got message: %q\n", string(msg.Data))
		msg.Ack()
		received++
		if received == 10 {
			cancel()
		}
	})
	if err != nil {
		return fmt.Errorf("Receive: %v", err)
	}
	return nil
}

Java

Antes de tentar essa amostra, siga as instruções de configuração do Java em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub Java.

Veja no GitHub Feedback


import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;

public class SubscribeAsyncExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";

    subscribeAsyncExample(projectId, subscriptionId);
  }

  public static void subscribeAsyncExample(String projectId, String subscriptionId) {
    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    // Instantiate an asynchronous message receiver.
    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          // Handle incoming message, then ack the received message.
          System.out.println("Id: " + message.getMessageId());
          System.out.println("Data: " + message.getData().toStringUtf8());
          consumer.ack();
        };

    Subscriber subscriber = null;
    try {
      subscriber = Subscriber.newBuilder(subscriptionName, receiver).build();
      // Start the subscriber.
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName.toString());
      // Allow the subscriber to run for 30s unless an unrecoverable error occurs.
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      // Shut down the subscriber after 30s. Stop receiving messages.
      subscriber.stopAsync();
    }
  }
}

Node.js

Antes de tentar essa amostra, siga as instruções de configuração do Node.js em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub Node.js.

Veja no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';
// const timeout = 60;

// Imports the Google Cloud client library
const {PubSub} = require('@google-cloud/pubsub');

// Creates a client; cache this for further use
const pubSubClient = new PubSub();

function listenForMessages() {
  // References an existing subscription
  const subscription = pubSubClient.subscription(subscriptionName);

  // Create an event handler to handle messages
  let messageCount = 0;
  const messageHandler = message => {
    console.log(`Received message ${message.id}:`);
    console.log(`\tData: ${message.data}`);
    console.log(`\tAttributes: ${message.attributes}`);
    messageCount += 1;

    // "Ack" (acknowledge receipt of) the message
    message.ack();
  };

  // Listen for new messages until timeout is hit
  subscription.on('message', messageHandler);

  setTimeout(() => {
    subscription.removeListener('message', messageHandler);
    console.log(`${messageCount} message(s) received.`);
  }, timeout * 1000);
}

listenForMessages();

Python

Antes de tentar esse exemplo, siga as instruções de configuração do Python em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub Python.

Veja no GitHub Feedback

from concurrent.futures import TimeoutError
from google.cloud import pubsub_v1

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"
# Number of seconds the subscriber should listen for messages
# timeout = 5.0

subscriber = pubsub_v1.SubscriberClient()
# The `subscription_path` method creates a fully qualified identifier
# in the form `projects/{project_id}/subscriptions/{subscription_id}`
subscription_path = subscriber.subscription_path(project_id, subscription_id)

def callback(message):
    print(f"Received {message}.")
    message.ack()

streaming_pull_future = subscriber.subscribe(subscription_path, callback=callback)
print(f"Listening for messages on {subscription_path}..\n")

# Wrap subscriber in a 'with' block to automatically call close() when done.
with subscriber:
    try:
        # When `timeout` is not set, result() will block indefinitely,
        # unless an exception is encountered first.
        streaming_pull_future.result(timeout=timeout)
    except TimeoutError:
        streaming_pull_future.cancel()  # Trigger the shutdown.
        streaming_pull_future.result()  # Block until the shutdown is complete.

Ruby

Antes de tentar esse exemplo, siga as instruções de configuração do Ruby em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub Ruby.

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
subscriber   = subscription.listen do |received_message|
  puts "Received message: #{received_message.data}"
  received_message.acknowledge!
end

subscriber.start
# Let the main thread sleep for 60 seconds so the thread for listening
# messages does not quit
sleep 60
subscriber.stop.wait!

Como processar atributos personalizados

Nesta amostra, você verá como enviar mensagens pull de maneira assíncrona e recuperar os atributos personalizados dos metadados:

C++

Veja no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
auto sample = [](pubsub::Subscriber subscriber) {
  return subscriber.Subscribe(
      [&](pubsub::Message const& m, pubsub::AckHandler h) {
        std::cout << "Received message with attributes:\n";
        for (auto const& kv : m.attributes()) {
          std::cout << "  " << kv.first << ": " << kv.second << "\n";
        }
        std::move(h).ack();
        PleaseIgnoreThisSimplifiesTestingTheSamples();
      });
};

C#

Veja no GitHub Feedback


using Google.Cloud.PubSub.V1;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading;
using System.Threading.Tasks;

public class PullMessagesWithCustomAttributesAsyncSample
{
    public async Task<List<PubsubMessage>> PullMessagesWithCustomAttributesAsync(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);

        SubscriberClient subscriber = await SubscriberClient.CreateAsync(subscriptionName);
        var messages = new List<PubsubMessage>();
        Task startTask = subscriber.StartAsync((PubsubMessage message, CancellationToken cancel) =>
        {
            messages.Add(message);
            string text = System.Text.Encoding.UTF8.GetString(message.Data.ToArray());
            Console.WriteLine($"Message {message.MessageId}: {text}");
            if (message.Attributes != null)
            {
                foreach (var attribute in message.Attributes)
                {
                    Console.WriteLine($"{attribute.Key} = {attribute.Value}");
                }
            }
            return Task.FromResult(acknowledge ? SubscriberClient.Reply.Ack : SubscriberClient.Reply.Nack);
        });
        // Run for 7 seconds.
        await Task.Delay(7000);
        await subscriber.StopAsync(CancellationToken.None);
        // Lets make sure that the start task finished successfully after the call to stop.
        await startTask;
        return messages;
    }
}

Go

Veja no GitHub Feedback

import (
	"context"
	"fmt"
	"io"
	"time"

	"cloud.google.com/go/pubsub"
)

func pullMsgsCustomAttributes(w io.Writer, projectID, subID string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}
	defer client.Close()

	sub := client.Subscription(subID)

	// Receive messages for 10 seconds.
	ctx, cancel := context.WithTimeout(ctx, 10*time.Second)
	defer cancel()

	// Create a channel to handle messages to as they come in.
	cm := make(chan *pubsub.Message)
	defer close(cm)

	// Handle individual messages in a goroutine.
	go func() {
		for msg := range cm {
			fmt.Fprintf(w, "Got message :%q\n", string(msg.Data))
			fmt.Fprintln(w, "Attributes:")
			for key, value := range msg.Attributes {
				fmt.Fprintf(w, "%s = %s", key, value)
			}
			msg.Ack()
		}
	}()

	// Receive blocks until the context is cancelled or an error occurs.
	err = sub.Receive(ctx, func(ctx context.Context, msg *pubsub.Message) {
		cm <- msg
	})
	if err != nil {
		return fmt.Errorf("Receive: %v", err)
	}

	return nil
}

Java

Veja no GitHub Feedback


import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;

public class SubscribeWithCustomAttributesExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";

    subscribeWithCustomAttributesExample(projectId, subscriptionId);
  }

  public static void subscribeWithCustomAttributesExample(String projectId, String subscriptionId) {
    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    // Instantiate an asynchronous message receiver.
    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          // Handle incoming message, then ack the received message.
          System.out.println("Id: " + message.getMessageId());
          System.out.println("Data: " + message.getData().toStringUtf8());
          // Print message attributes.
          message
              .getAttributesMap()
              .forEach((key, value) -> System.out.println(key + " = " + value));
          consumer.ack();
        };

    Subscriber subscriber = null;
    try {
      subscriber = Subscriber.newBuilder(subscriptionName, receiver).build();
      // Start the subscriber.
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName.toString());
      // Allow the subscriber to run for 30s unless an unrecoverable error occurs.
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      // Shut down the subscriber after 30s. Stop receiving messages.
      subscriber.stopAsync();
    }
  }
}

Node.js

Veja no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';
// const timeout = 60;

// Imports the Google Cloud client library
const {PubSub} = require('@google-cloud/pubsub');

// Creates a client; cache this for further use
const pubSubClient = new PubSub();

async function listenWithCustomAttributes() {
  // References an existing subscription, e.g. "my-subscription"
  const subscription = pubSubClient.subscription(subscriptionName);

  // Create an event handler to handle messages
  const messageHandler = message => {
    console.log(
      `Received message: id ${message.id}, data ${
        message.data
      }, attributes: ${JSON.stringify(message.attributes)}`
    );

    // "Ack" (acknowledge receipt of) the message
    message.ack();
  };

  // Listen for new messages until timeout is hit
  subscription.on('message', messageHandler);
  setTimeout(() => {
    subscription.removeListener('message', messageHandler);
  }, timeout * 1000);
}

listenWithCustomAttributes();

Python

Veja no GitHub Feedback

from concurrent.futures import TimeoutError
from google.cloud import pubsub_v1

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"
# Number of seconds the subscriber should listen for messages
# timeout = 5.0

subscriber = pubsub_v1.SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

def callback(message):
    print(f"Received {message.data}.")
    if message.attributes:
        print("Attributes:")
        for key in message.attributes:
            value = message.attributes.get(key)
            print(f"{key}: {value}")
    message.ack()

streaming_pull_future = subscriber.subscribe(subscription_path, callback=callback)
print(f"Listening for messages on {subscription_path}..\n")

# Wrap subscriber in a 'with' block to automatically call close() when done.
with subscriber:
    try:
        # When `timeout` is not set, result() will block indefinitely,
        # unless an exception is encountered first.
        streaming_pull_future.result(timeout=timeout)
    except TimeoutError:
        streaming_pull_future.cancel()  # Trigger the shutdown.
        streaming_pull_future.result()  # Block until the shutdown is complete.

Ruby

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
subscriber   = subscription.listen do |received_message|
  puts "Received message: #{received_message.data}"
  unless received_message.attributes.empty?
    puts "Attributes:"
    received_message.attributes.each do |key, value|
      puts "#{key}: #{value}"
    end
  end
  received_message.acknowledge!
end

subscriber.start
# Let the main thread sleep for 60 seconds so the thread for listening
# messages does not quit
sleep 60
subscriber.stop.wait!

Escuta de erros

Esta amostra exibe como lidar com erros que ocorrem durante a assinatura nas mensagens:

C++

Veja no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
using google::cloud::future;
auto sample = [](pubsub::Subscriber subscriber) {
  return subscriber
      .Subscribe([&](pubsub::Message const& m, pubsub::AckHandler h) {
        std::cout << "Received message " << m << "\n";
        std::move(h).ack();
        PleaseIgnoreThisSimplifiesTestingTheSamples();
      })
      // Setup an error handler for the subscription session
      .then([](future<google::cloud::Status> f) {
        std::cout << "Subscription session result: " << f.get() << "\n";
      });
};

Go

Veja no GitHub Feedback

import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/pubsub"
)

func pullMsgsError(w io.Writer, projectID, subID string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}
	defer client.Close()

	// If the service returns a non-retryable error, Receive returns that error after
	// all of the outstanding calls to the handler have returned.
	err = client.Subscription(subID).Receive(ctx, func(ctx context.Context, msg *pubsub.Message) {
		fmt.Fprintf(w, "Got message: %q\n", string(msg.Data))
		msg.Ack()
	})
	if err != nil {
		return fmt.Errorf("Receive: %v", err)
	}
	return nil
}

Java

Veja no GitHub Feedback


import com.google.api.gax.core.ExecutorProvider;
import com.google.api.gax.core.InstantiatingExecutorProvider;
import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.common.util.concurrent.MoreExecutors;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;

public class SubscribeWithErrorListenerExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";

    subscribeWithErrorListenerExample(projectId, subscriptionId);
  }

  public static void subscribeWithErrorListenerExample(String projectId, String subscriptionId) {
    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    // Instantiate an asynchronous message receiver.
    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          // Handle incoming message, then ack the received message.
          System.out.println("Id: " + message.getMessageId());
          System.out.println("Data: " + message.getData().toStringUtf8());
          consumer.ack();
        };

    Subscriber subscriber = null;
    try {
      // Provides an executor service for processing messages.
      ExecutorProvider executorProvider =
          InstantiatingExecutorProvider.newBuilder().setExecutorThreadCount(4).build();

      subscriber =
          Subscriber.newBuilder(subscriptionName, receiver)
              .setExecutorProvider(executorProvider)
              .build();

      // Listen for unrecoverable failures. Rebuild a subscriber and restart subscribing
      // when the current subscriber encounters permanent errors.
      subscriber.addListener(
          new Subscriber.Listener() {
            public void failed(Subscriber.State from, Throwable failure) {
              System.out.println(failure.getStackTrace());
              if (!executorProvider.getExecutor().isShutdown()) {
                subscribeWithErrorListenerExample(projectId, subscriptionId);
              }
            }
          },
          MoreExecutors.directExecutor());

      // Start the subscriber.
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName.toString());
      // Allow the subscriber to run for 30s unless an unrecoverable error occurs.
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      // Shut down the subscriber after 30s. Stop receiving messages.
      subscriber.stopAsync();
    }
  }
}

Node.js

Veja no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';
// const timeout = 10;

// Imports the Google Cloud client library
const {PubSub} = require('@google-cloud/pubsub');

// Creates a client; cache this for further use
const pubSubClient = new PubSub();

function listenForErrors() {
  // References an existing subscription
  const subscription = pubSubClient.subscription(subscriptionName);

  // Create an event handler to handle messages
  const messageHandler = function (message) {
    // Do something with the message
    console.log(`Message: ${message}`);

    // "Ack" (acknowledge receipt of) the message
    message.ack();
  };

  // Create an event handler to handle errors
  const errorHandler = function (error) {
    // Do something with the error
    console.error(`ERROR: ${error}`);
    throw error;
  };

  // Listen for new messages/errors until timeout is hit
  subscription.on('message', messageHandler);
  subscription.on('error', errorHandler);

  setTimeout(() => {
    subscription.removeListener('message', messageHandler);
    subscription.removeListener('error', errorHandler);
  }, timeout * 1000);
}

listenForErrors();

Python

Veja no GitHub Feedback

from google.cloud import pubsub_v1

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"
# Number of seconds the subscriber should listen for messages
# timeout = 5.0

subscriber = pubsub_v1.SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

def callback(message):
    print(f"Received {message}.")
    message.ack()

streaming_pull_future = subscriber.subscribe(subscription_path, callback=callback)
print(f"Listening for messages on {subscription_path}..\n")

# Wrap subscriber in a 'with' block to automatically call close() when done.
with subscriber:
    # When `timeout` is not set, result() will block indefinitely,
    # unless an exception is encountered first.
    try:
        streaming_pull_future.result(timeout=timeout)
    except Exception as e:
        print(
            f"Listening for messages on {subscription_path} threw an exception: {e}."
        )
        streaming_pull_future.cancel()  # Trigger the shutdown.
        streaming_pull_future.result()  # Block until the shutdown is complete.

Ruby

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
subscriber   = subscription.listen do |received_message|
  puts "Received message: #{received_message.data}"
  received_message.acknowledge!
end
# Propagate expection from child threads to the main thread as soon as it is
# raised. Exceptions happened in the callback thread are collected in the
# callback thread pool and do not propagate to the main thread
Thread.abort_on_exception = true

begin
  subscriber.start
  # Let the main thread sleep for 60 seconds so the thread for listening
  # messages does not quit
  sleep 60
  subscriber.stop.wait!
rescue Exception => e
  puts "Exception #{e.inspect}: #{e.message}"
  raise "Stopped listening for messages."
end

Controle de fluxo de mensagens

O cliente assinante pode processar e reconhecer as mensagens mais lentamente do que o Pub/Sub as envia. Nesse caso:

É possível que um cliente tenha um backlog de mensagens por não ter capacidade para processar o volume de mensagens recebidas, mas outro cliente na rede tem. Este pode reduzir o backlog da assinatura, mas não consegue fazê-lo porque o primeiro cliente mantém um lease nas mensagens recebidas. Isso reduz a taxa de processamento geral, porque as mensagens ficam presas no primeiro cliente.
Como a biblioteca de cliente estende repetidamente o prazo de confirmação para as mensagens acumuladas, elas continuam a consumir recursos de memória, CPU e largura de banda. Dessa forma, o cliente assinante pode ficar sem recursos (como memória). Isso pode afetar negativamente a capacidade e a latência das mensagens de processamento.

Para atenuar os problemas acima, controle a taxa com que o assinante recebe as mensagens com os recursos de controle de fluxo do assinante. Esses recursos de controle de fluxo são ilustrados nas amostras a seguir:

C++

Veja no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
using google::cloud::future;
using google::cloud::StatusOr;
auto sample = [](std::string project_id, std::string subscription_id) {
  // Change the flow control watermarks, by default the client library uses
  // 0 and 1,000 for the message count watermarks, and 0 and 10MiB for the
  // size watermarks. Recall that the library stops requesting messages if
  // any of the high watermarks are reached, and the library resumes
  // requesting messages when *both* low watermarks are reached.
  auto constexpr kMiB = 1024 * 1024L;
  auto subscriber = pubsub::Subscriber(pubsub::MakeSubscriberConnection(
      pubsub::Subscription(std::move(project_id), std::move(subscription_id)),
      pubsub::SubscriberOptions{}
          .set_max_outstanding_messages(1000)
          .set_max_outstanding_bytes(8 * kMiB)));

  auto session = subscriber.Subscribe(
      [](pubsub::Message const& m, pubsub::AckHandler h) {
        std::move(h).ack();
        std::cout << "Received message " << m << "\n";
        PleaseIgnoreThisSimplifiesTestingTheSamples();
      });
  return std::make_pair(subscriber, std::move(session));
};

C#

Veja no GitHub Feedback


using Google.Api.Gax;
using Google.Cloud.PubSub.V1;
using System;
using System.Linq;
using System.Text;
using System.Threading;
using System.Threading.Tasks;

public class PullMessagesWithFlowControlAsyncSample
{
    public async Task<int> PullMessagesWithFlowControlAsync(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);
        int messageCount = 0;
        SubscriberClient subscriber = await SubscriberClient.CreateAsync(subscriptionName,
            settings: new SubscriberClient.Settings()
            {
                AckExtensionWindow = TimeSpan.FromSeconds(4),
                AckDeadline = TimeSpan.FromSeconds(10),
                FlowControlSettings = new FlowControlSettings(maxOutstandingElementCount: 100, maxOutstandingByteCount: 10240)
            });
        // SubscriberClient runs your message handle function on multiple
        // threads to maximize throughput.
        Task startTask = subscriber.StartAsync((PubsubMessage message, CancellationToken cancel) =>
        {
            string text = System.Text.Encoding.UTF8.GetString(message.Data.ToArray());
            Console.WriteLine($"Message {message.MessageId}: {text}");
            Interlocked.Increment(ref messageCount);
            return Task.FromResult(acknowledge ? SubscriberClient.Reply.Ack : SubscriberClient.Reply.Nack);
        });
        // Run for 5 seconds.
        await Task.Delay(5000);
        await subscriber.StopAsync(CancellationToken.None);
        // Lets make sure that the start task finished successfully after the call to stop.
        await startTask;
        return messageCount;
    }
}

Go

Veja no GitHub Feedback

import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/pubsub"
)

func pullMsgsSettings(w io.Writer, projectID, subID string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}
	defer client.Close()

	sub := client.Subscription(subID)
	sub.ReceiveSettings.Synchronous = true
	// MaxOutstandingMessages is the maximum number of unprocessed messages the
	// client will pull from the server before pausing.
	//
	// This is only guaranteed when ReceiveSettings.Synchronous is set to true.
	// When Synchronous is set to false, the StreamingPull RPC is used which
	// can pull a single large batch of messages at once that is greater than
	// MaxOustandingMessages before pausing. For more info, see
	// https://cloud.google.com/pubsub/docs/pull#streamingpull_dealing_with_large_backlogs_of_small_messages.
	sub.ReceiveSettings.MaxOutstandingMessages = 10
	// MaxOutstandingBytes is the maximum size of unprocessed messages,
	// that the client will pull from the server before pausing. Similar
	// to MaxOutstandingMessages, this may be exceeded with a large batch
	// of messages since we cannot control the size of a batch of messages
	// from the server (even with the synchronous Pull RPC).
	sub.ReceiveSettings.MaxOutstandingBytes = 1e10
	err = sub.Receive(ctx, func(ctx context.Context, msg *pubsub.Message) {
		fmt.Fprintf(w, "Got message: %q\n", string(msg.Data))
		msg.Ack()
	})
	if err != nil {
		return fmt.Errorf("Receive: %v", err)
	}
	return nil
}

Java

Veja no GitHub Feedback


import com.google.api.gax.batching.FlowControlSettings;
import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;

public class SubscribeWithFlowControlSettingsExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";

    subscribeWithFlowControlSettingsExample(projectId, subscriptionId);
  }

  public static void subscribeWithFlowControlSettingsExample(
      String projectId, String subscriptionId) {
    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    // Instantiate an asynchronous message receiver.
    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          // Handle incoming message, then ack the received message.
          System.out.println("Id: " + message.getMessageId());
          System.out.println("Data: " + message.getData().toStringUtf8());
          consumer.ack();
        };

    Subscriber subscriber = null;

    // The subscriber will pause the message stream and stop receiving more messsages from the
    // server if any one of the conditions is met.
    FlowControlSettings flowControlSettings =
        FlowControlSettings.newBuilder()
            // 1,000 outstanding messages. Must be >0. It controls the maximum number of messages
            // the subscriber receives before pausing the message stream.
            .setMaxOutstandingElementCount(1000L)
            // 100 MiB. Must be >0. It controls the maximum size of messages the subscriber
            // receives before pausing the message stream.
            .setMaxOutstandingRequestBytes(100L * 1024L * 1024L)
            .build();

    try {
      subscriber =
          Subscriber.newBuilder(subscriptionName, receiver)
              .setFlowControlSettings(flowControlSettings)
              .build();

      // Start the subscriber.
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName.toString());
      // Allow the subscriber to run for 30s unless an unrecoverable error occurs.
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      // Shut down the subscriber after 30s. Stop receiving messages.
      subscriber.stopAsync();
    }
  }
}

Node.js

Veja no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';
// const maxInProgress = 5;
// const timeout = 10;

// Imports the Google Cloud client library
const {PubSub} = require('@google-cloud/pubsub');

// Creates a client; cache this for further use
const pubSubClient = new PubSub();

async function subscribeWithFlowControlSettings() {
  const subscriberOptions = {
    flowControl: {
      maxMessages: maxInProgress,
    },
  };

  // References an existing subscription.
  // Note that flow control settings are not persistent across subscribers.
  const subscription = pubSubClient.subscription(
    subscriptionName,
    subscriberOptions
  );

  console.log(
    `Subscriber to subscription ${subscription.name} is ready to receive messages at a controlled volume of ${maxInProgress} messages.`
  );

  const messageHandler = message => {
    console.log(`Received message: ${message.id}`);
    console.log(`\tData: ${message.data}`);
    console.log(`\tAttributes: ${message.attributes}`);

    // "Ack" (acknowledge receipt of) the message
    message.ack();
  };

  subscription.on('message', messageHandler);

  setTimeout(() => {
    subscription.close();
  }, timeout * 1000);
}

Python

Veja no GitHub Feedback

from concurrent.futures import TimeoutError
from google.cloud import pubsub_v1

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"
# Number of seconds the subscriber should listen for messages
# timeout = 5.0

subscriber = pubsub_v1.SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

def callback(message):
    print(f"Received {message.data}.")
    message.ack()

# Limit the subscriber to only have ten outstanding messages at a time.
flow_control = pubsub_v1.types.FlowControl(max_messages=10)

streaming_pull_future = subscriber.subscribe(
    subscription_path, callback=callback, flow_control=flow_control
)
print(f"Listening for messages on {subscription_path}..\n")

# Wrap subscriber in a 'with' block to automatically call close() when done.
with subscriber:
    try:
        # When `timeout` is not set, result() will block indefinitely,
        # unless an exception is encountered first.
        streaming_pull_future.result(timeout=timeout)
    except TimeoutError:
        streaming_pull_future.cancel()  # Trigger the shutdown.
        streaming_pull_future.result()  # Block until the shutdown is complete.

Ruby

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
subscriber   = subscription.listen inventory: 10 do |received_message|
  puts "Received message: #{received_message.data}"
  received_message.acknowledge!
end

subscriber.start
# Let the main thread sleep for 60 seconds so the thread for listening
# messages does not quit
sleep 60
subscriber.stop.wait!

De modo mais geral, a necessidade do controle de fluxo indica que as mensagens são publicadas com uma taxa maior do que são consumidas. Se esse for um estado constante, em vez de um pico transitório no volume de mensagens, pense em aumentar o número de instâncias do cliente assinante.

Controle de simultaneidade

A compatibilidade com a simultaneidade depende da linguagem de programação. Para implementações de linguagem compatíveis com linhas de execução paralelas, como Java e Go, as bibliotecas de cliente são a escolha padrão para o número de linhas de execução. Essa escolha pode não ser ideal para o aplicativo. Por exemplo, caso ache que o aplicativo assinante não está acompanhando o volume de mensagens recebidas nem está ligado à CPU, aumente a contagem de linhas de execução. Para operações de processamento de mensagens intensivas da CPU, diminua o número de linhas de execução.

A amostra seguir ilustra como controlar a simultaneidade em um assinante:

C++

Veja no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
using google::cloud::future;
using google::cloud::StatusOr;
auto sample = [](std::string project_id, std::string subscription_id) {
  // Create a subscriber with 16 threads handling I/O work, by default the
  // library creates `std::thread::hardware_concurrency()` threads.
  auto subscriber = pubsub::Subscriber(pubsub::MakeSubscriberConnection(
      pubsub::Subscription(std::move(project_id), std::move(subscription_id)),
      pubsub::SubscriberOptions{}.set_max_concurrency(8),
      pubsub::ConnectionOptions{}.set_background_thread_pool_size(16)));

  // Create a subscription where up to 8 messages are handled concurrently. By
  // default the library uses `std::thread::hardware_concurrency()` as the
  // maximum number of concurrent callbacks.
  auto session = subscriber.Subscribe(
      [](pubsub::Message const& m, pubsub::AckHandler h) {
        // This handler executes in the I/O threads, applications could use,
        // std::async(), a thread-pool, or any other mechanism to transfer the
        // execution to other threads.
        std::cout << "Received message " << m << "\n";
        std::move(h).ack();
        PleaseIgnoreThisSimplifiesTestingTheSamples();
      });
  return std::make_pair(subscriber, std::move(session));
};

Go

Veja no GitHub Feedback

import (
	"context"
	"fmt"
	"io"
	"runtime"
	"sync/atomic"
	"time"

	"cloud.google.com/go/pubsub"
)

func pullMsgsConcurrenyControl(w io.Writer, projectID, subID string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}
	defer client.Close()

	sub := client.Subscription(subID)
	// Must set ReceiveSettings.Synchronous to false (or leave as default) to enable
	// concurrency settings. Otherwise, NumGoroutines will be set to 1.
	sub.ReceiveSettings.Synchronous = false
	// NumGoroutines is the number of goroutines sub.Receive will spawn to pull
	// messages concurrently.
	sub.ReceiveSettings.NumGoroutines = runtime.NumCPU()

	// Receive messages for 10 seconds.
	ctx, cancel := context.WithTimeout(ctx, 10*time.Second)
	defer cancel()

	var counter int32

	// Receive blocks until the context is cancelled or an error occurs.
	err = sub.Receive(ctx, func(_ context.Context, msg *pubsub.Message) {
		// The message handler passed to Receive may be called concurrently
		// so it's okay to process the messages concurrently but make sure
		// to synchronize access to shared memory.
		atomic.AddInt32(&counter, 1)
		msg.Ack()
	})
	if err != nil {
		return fmt.Errorf("pubsub: Receive returned error: %v", err)
	}
	fmt.Fprintf(w, "Received %d messages\n", counter)

	return nil
}

Java

Veja no GitHub Feedback


import com.google.api.gax.core.ExecutorProvider;
import com.google.api.gax.core.InstantiatingExecutorProvider;
import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;

public class SubscribeWithConcurrencyControlExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";

    subscribeWithConcurrencyControlExample(projectId, subscriptionId);
  }

  public static void subscribeWithConcurrencyControlExample(
      String projectId, String subscriptionId) {
    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    // Instantiate an asynchronous message receiver.
    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          // Handle incoming message, then ack the received message.
          System.out.println("Id: " + message.getMessageId());
          System.out.println("Data: " + message.getData().toStringUtf8());
          consumer.ack();
        };

    Subscriber subscriber = null;
    try {
      // Provides an executor service for processing messages. The default `executorProvider` used
      // by the subscriber has a default thread count of 5.
      ExecutorProvider executorProvider =
          InstantiatingExecutorProvider.newBuilder().setExecutorThreadCount(4).build();

      // `setParallelPullCount` determines how many StreamingPull streams the subscriber will open
      // to receive message. It defaults to 1. `setExecutorProvider` configures an executor for the
      // subscriber to process messages. Here, the subscriber is configured to open 2 streams for
      // receiving messages, each stream creates a new executor with 4 threads to help process the
      // message callbacks. In total 2x4=8 threads are used for message processing.
      subscriber =
          Subscriber.newBuilder(subscriptionName, receiver)
              .setParallelPullCount(2)
              .setExecutorProvider(executorProvider)
              .build();

      // Start the subscriber.
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName.toString());
      // Allow the subscriber to run for 30s unless an unrecoverable error occurs.
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      // Shut down the subscriber after 30s. Stop receiving messages.
      subscriber.stopAsync();
    }
  }
}

Ruby

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
# Use 2 threads for streaming, 4 threads for executing callbacks and 2 threads
# for sending acknowledgements and/or delays
subscriber   = subscription.listen streams: 2, threads: {
  callback: 4,
  push:     2
} do |received_message|
  puts "Received message: #{received_message.data}"
  received_message.acknowledge!
end

subscriber.start
# Let the main thread sleep for 60 seconds so the thread for listening
# messages does not quit
sleep 60
subscriber.stop.wait!

A compatibilidade com a simultaneidade depende da linguagem de programação. Consulte a documentação de referência de APIs para mais informações.

Como usar o esquema

Estas amostras mostram como processar mensagens ao se inscrever em mensagens em tópicos configurados com esquema. Para saber mais, consulte Como criar e gerenciar esquemas.

C++

Avro

Ver no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
using google::cloud::future;
using google::cloud::StatusOr;
return [](pubsub::Subscriber subscriber) {
  auto session = subscriber.Subscribe(
      [](pubsub::Message const& m, pubsub::AckHandler h) {
        std::cout << "Message contents: " << m.data() << "\n";
        std::move(h).ack();
      });
  return session;
}

Proto

Ver no GitHub Feedback

namespace pubsub = google::cloud::pubsub;
using google::cloud::future;
using google::cloud::StatusOr;
return [](pubsub::Subscriber subscriber) {
  auto session = subscriber.Subscribe(
      [](pubsub::Message const& m, pubsub::AckHandler h) {
        google::cloud::pubsub::samples::State state;
        state.ParseFromString(m.data());
        std::cout << "Message contents: " << state.DebugString() << "\n";
        std::move(h).ack();
      });
  return session;
}

C#

Avro

Ver no GitHub Feedback


using Avro.IO;
using Avro.Specific;
using Google.Api.Gax;
using Google.Cloud.PubSub.V1;
using Newtonsoft.Json;
using System;
using System.IO;
using System.Threading;
using System.Threading.Tasks;

public class PullAvroMessagesAsyncSample
{
    public async Task<int> PullAvroMessagesAsync(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);
        int messageCount = 0;
        SubscriberClient subscriber = await SubscriberClient.CreateAsync(subscriptionName,
            settings: new SubscriberClient.Settings()
            {
                AckExtensionWindow = TimeSpan.FromSeconds(4),
                AckDeadline = TimeSpan.FromSeconds(10),
                FlowControlSettings = new FlowControlSettings(maxOutstandingElementCount: 100, maxOutstandingByteCount: 10240)
            });
        // SubscriberClient runs your message handle function on multiple
        // threads to maximize throughput.
        Task startTask = subscriber.StartAsync((PubsubMessage message, CancellationToken cancel) =>
        {
            string encoding = message.Attributes["googclient_schemaencoding"];
            AvroUtilities.State state = new AvroUtilities.State();
            switch (encoding)
            {
                case "BINARY":
                    using (var ms = new MemoryStream(message.Data.ToByteArray()))
                    {
                        var decoder = new BinaryDecoder(ms);
                        var reader = new SpecificDefaultReader(state.Schema, state.Schema);
                        reader.Read<AvroUtilities.State>(state, decoder);
                    }
                    break;
                case "JSON":
                    state = JsonConvert.DeserializeObject<AvroUtilities.State>(message.Data.ToStringUtf8());
                    break;
                default:
                    Console.WriteLine($"Encoding not provided in message.");
                    break;
            }
            Console.WriteLine($"Message {message.MessageId}: {state}");
            Interlocked.Increment(ref messageCount);
            return Task.FromResult(acknowledge ? SubscriberClient.Reply.Ack : SubscriberClient.Reply.Nack);
        });
        // Run for 5 seconds.
        await Task.Delay(5000);
        await subscriber.StopAsync(CancellationToken.None);
        // Lets make sure that the start task finished successfully after the call to stop.
        await startTask;
        return messageCount;
    }
}

Proto

Ver no GitHub Feedback


using Google.Api.Gax;
using Google.Cloud.PubSub.V1;
using System;
using System.Threading;
using System.Threading.Tasks;

public class PullProtoMessagesAsyncSample
{
    public async Task<int> PullProtoMessagesAsync(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);
        int messageCount = 0;
        SubscriberClient subscriber = await SubscriberClient.CreateAsync(subscriptionName,
            settings: new SubscriberClient.Settings()
            {
                AckExtensionWindow = TimeSpan.FromSeconds(4),
                AckDeadline = TimeSpan.FromSeconds(10),
                FlowControlSettings = new FlowControlSettings(maxOutstandingElementCount: 100, maxOutstandingByteCount: 10240)
            });
        // SubscriberClient runs your message handle function on multiple
        // threads to maximize throughput.
        Task startTask = subscriber.StartAsync((PubsubMessage message, CancellationToken cancel) =>
        {
            string encoding = message.Attributes["googclient_schemaencoding"];
            Utilities.State state = null;
            switch (encoding)
            {
                case "BINARY":
                    state = Utilities.State.Parser.ParseFrom(message.Data.ToByteArray());
                    break;
                case "JSON":
                    state = Utilities.State.Parser.ParseJson(message.Data.ToStringUtf8());
                    break;
                default:
                    Console.WriteLine($"Encoding not provided in message.");
                    break;
            }
            Console.WriteLine($"Message {message.MessageId}: {state}");
            Interlocked.Increment(ref messageCount);
            return Task.FromResult(acknowledge ? SubscriberClient.Reply.Ack : SubscriberClient.Reply.Nack);
        });
        // Run for 5 seconds.
        await Task.Delay(5000);
        await subscriber.StopAsync(CancellationToken.None);
        // Lets make sure that the start task finished successfully after the call to stop.
        await startTask;
        return messageCount;
    }
}

Go

Avro

Ver no GitHub Feedback

import (
	"context"
	"fmt"
	"io"
	"io/ioutil"
	"sync"
	"time"

	"cloud.google.com/go/pubsub"
	"github.com/linkedin/goavro/v2"
)

func subscribeWithAvroSchema(w io.Writer, projectID, subID, avscFile string) error {
	// projectID := "my-project-id"
	// topicID := "my-topic"
	// avscFile = "path/to/an/avro/schema/file(.avsc)/formatted/in/json"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}

	avroSchema, err := ioutil.ReadFile(avscFile)
	if err != nil {
		return fmt.Errorf("ioutil.ReadFile err: %v", err)
	}
	codec, err := goavro.NewCodec(string(avroSchema))
	if err != nil {
		return fmt.Errorf("goavro.NewCodec err: %v", err)
	}

	sub := client.Subscription(subID)
	ctx2, cancel := context.WithTimeout(ctx, 10*time.Second)
	defer cancel()

	var mu sync.Mutex
	sub.Receive(ctx2, func(ctx context.Context, msg *pubsub.Message) {
		mu.Lock()
		defer mu.Unlock()
		encoding := msg.Attributes["googclient_schemaencoding"]

		if encoding == "BINARY" {
			data, _, err := codec.NativeFromBinary(msg.Data)
			if err != nil {
				fmt.Fprintf(w, "codec.NativeFromBinary err: %v", err)
				return
			}
			fmt.Printf("Received a binary-encoded message:\n%#v", data)
		} else if encoding == "JSON" {
			data, _, err := codec.NativeFromTextual(msg.Data)
			if err != nil {
				fmt.Fprintf(w, "codec.NativeFromTextual err: %v", err)
				return
			}
			fmt.Fprintf(w, "Received a JSON-encoded message:\n%#v", data)
		} else {
			fmt.Fprintf(w, "invalid encoding: %s", encoding)
			return
		}
		msg.Ack()
	})
	return nil
}

Proto

Ver no GitHub Feedback

import (
	"context"
	"fmt"
	"io"
	"sync"
	"time"

	"cloud.google.com/go/pubsub"
	statepb "github.com/GoogleCloudPlatform/golang-samples/internal/pubsub/schemas"
	"google.golang.org/protobuf/encoding/protojson"
	"google.golang.org/protobuf/proto"
)

func subscribeWithProtoSchema(w io.Writer, projectID, subID, protoFile string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	// protoFile = "path/to/a/proto/schema/file(.proto)/formatted/in/protocol/buffers"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}

	// Create an instance of the message to be decoded (a single U.S. state).
	state := &statepb.State{}

	sub := client.Subscription(subID)
	ctx2, cancel := context.WithTimeout(ctx, 10*time.Second)
	defer cancel()

	var mu sync.Mutex
	sub.Receive(ctx2, func(ctx context.Context, msg *pubsub.Message) {
		mu.Lock()
		defer mu.Unlock()
		encoding := msg.Attributes["googclient_schemaencoding"]

		if encoding == "BINARY" {
			if err := proto.Unmarshal(msg.Data, state); err != nil {
				fmt.Fprintf(w, "proto.Unmarshal err: %v", err)
				return
			}
			fmt.Printf("Received a binary-encoded message:\n%#v", state)
		} else if encoding == "JSON" {
			if err := protojson.Unmarshal(msg.Data, state); err != nil {
				fmt.Fprintf(w, "proto.Unmarshal err: %v", err)
				return
			}
			fmt.Fprintf(w, "Received a JSON-encoded message:\n%#v", state)
		} else {
			fmt.Fprintf(w, "invalid encoding: %s", encoding)
		}
		msg.Ack()
	})
	return nil
}

Java

Avro

Ver no GitHub Feedback


import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.protobuf.ByteString;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.io.ByteArrayInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;
import org.apache.avro.io.Decoder;
import org.apache.avro.io.DecoderFactory;
import org.apache.avro.specific.SpecificDatumReader;
import utilities.State;

public class SubscribeWithAvroSchemaExample {

  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    // Use an existing subscription.
    String subscriptionId = "your-subscription-id";

    subscribeWithAvroSchemaExample(projectId, subscriptionId);
  }

  public static void subscribeWithAvroSchemaExample(String projectId, String subscriptionId) {

    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    // Prepare a reader for the encoded Avro records.
    SpecificDatumReader<State> reader = new SpecificDatumReader<>(State.getClassSchema());

    // Instantiate an asynchronous message receiver.
    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          ByteString data = message.getData();

          // Get the schema encoding type.
          String encoding = message.getAttributesMap().get("googclient_schemaencoding");

          // Send the message data to a byte[] input stream.
          InputStream inputStream = new ByteArrayInputStream(data.toByteArray());

          Decoder decoder = null;

          // Prepare an appropriate decoder for the message data in the input stream
          // based on the schema encoding type.
          block:
          try {
            switch (encoding) {
              case "BINARY":
                decoder = DecoderFactory.get().directBinaryDecoder(inputStream, /*reuse=*/ null);
                System.out.println("Receiving a binary-encoded message:");
                break;
              case "JSON":
                decoder = DecoderFactory.get().jsonDecoder(State.getClassSchema(), inputStream);
                System.out.println("Receiving a JSON-encoded message:");
                break;
              default:
                break block;
            }

            // Obtain an object of the generated Avro class using the decoder.
            State state = reader.read(null, decoder);
            System.out.println(state.getName() + " is abbreviated as " + state.getPostAbbr());

          } catch (IOException e) {
            System.err.println(e);
          }

          // Ack the message.
          consumer.ack();
        };

    Subscriber subscriber = null;
    try {
      subscriber = Subscriber.newBuilder(subscriptionName, receiver).build();
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName.toString());
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      subscriber.stopAsync();
    }
  }
}

Buffer de protocolo

Ver no GitHub Feedback


import com.google.cloud.pubsub.v1.AckReplyConsumer;
import com.google.cloud.pubsub.v1.MessageReceiver;
import com.google.cloud.pubsub.v1.Subscriber;
import com.google.protobuf.ByteString;
import com.google.protobuf.InvalidProtocolBufferException;
import com.google.protobuf.util.JsonFormat;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PubsubMessage;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;
import utilities.StateProto.State;

public class SubscribeWithProtoSchemaExample {

  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    // Use an existing subscription.
    String subscriptionId = "your-subscription-id";

    subscribeWithProtoSchemaExample(projectId, subscriptionId);
  }

  public static void subscribeWithProtoSchemaExample(String projectId, String subscriptionId) {

    ProjectSubscriptionName subscriptionName =
        ProjectSubscriptionName.of(projectId, subscriptionId);

    MessageReceiver receiver =
        (PubsubMessage message, AckReplyConsumer consumer) -> {
          ByteString data = message.getData();

          // Get the schema encoding type.
          String encoding = message.getAttributesMap().get("googclient_schemaencoding");

          block:
          try {
            switch (encoding) {
              case "BINARY":
                // Obtain an object of the generated proto class.
                State state = State.parseFrom(data);
                System.out.println("Received a BINARY-formatted message: " + state);
                break;

              case "JSON":
                State.Builder stateBuilder = State.newBuilder();
                JsonFormat.parser().merge(data.toStringUtf8(), stateBuilder);
                System.out.println("Received a JSON-formatted message:" + stateBuilder.build());
                break;

              default:
                break block;
            }
          } catch (InvalidProtocolBufferException e) {
            e.printStackTrace();
          }

          consumer.ack();
          System.out.println("Ack'ed the message");
        };

    // Create subscriber client.
    Subscriber subscriber = Subscriber.newBuilder(subscriptionName, receiver).build();

    try {
      subscriber.startAsync().awaitRunning();
      System.out.printf("Listening for messages on %s:\n", subscriptionName);
      subscriber.awaitTerminated(30, TimeUnit.SECONDS);
    } catch (TimeoutException timeoutException) {
      subscriber.stopAsync();
    }
  }
}

Node.js

Avro

Ver no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';
// const timeout = 60;

// Imports the Google Cloud client library
const {PubSub, Schema, Encodings} = require('@google-cloud/pubsub');

// Node FS library, to load definitions
const fs = require('fs');

// And the Apache Avro library
const avro = require('avro-js');

// Creates a client; cache this for further use
const pubSubClient = new PubSub();

function listenForAvroRecords(subscriptionName, timeout) {
  // References an existing subscription
  const subscription = pubSubClient.subscription(subscriptionName);

  // Make an encoder using the official avro-js library.
  const definition = fs
    .readFileSync('system-test/fixtures/provinces.avsc')
    .toString();
  const type = avro.parse(definition);

  // Create an event handler to handle messages
  let messageCount = 0;
  const messageHandler = async message => {
    // "Ack" (acknowledge receipt of) the message
    message.ack();

    // Get the schema metadata from the message.
    const schemaMetadata = Schema.metadataFromMessage(message.attributes);

    let result;
    switch (schemaMetadata.encoding) {
      case Encodings.Binary:
        result = type.fromBuffer(message.data);
        break;
      case Encodings.Json:
        result = type.fromString(message.data.toString());
        break;
      default:
        console.log(`Unknown schema encoding: ${schemaMetadata.encoding}`);
        break;
    }

    console.log(`Received message ${message.id}:`);
    console.log(`\tData: ${JSON.stringify(result, null, 4)}`);
    console.log(`\tAttributes: ${message.attributes}`);
    messageCount += 1;
  };

  // Listen for new messages until timeout is hit
  subscription.on('message', messageHandler);

  setTimeout(() => {
    subscription.removeListener('message', messageHandler);
    console.log(`${messageCount} message(s) received.`);
  }, timeout * 1000);
}

Buffer de protocolo

Ver no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';
// const timeout = 60;

// Imports the Google Cloud client library
const {PubSub, Schema, Encodings} = require('@google-cloud/pubsub');

// And the protobufjs library
const protobuf = require('protobufjs');

// Creates a client; cache this for further use
const pubSubClient = new PubSub();

async function listenForProtobufMessages(subscriptionName, timeout) {
  // References an existing subscription
  const subscription = pubSubClient.subscription(subscriptionName);

  // Make an decoder using the protobufjs library.
  //
  // Since we're providing the test message for a specific schema here, we'll
  // also code in the path to a sample proto definition.
  const root = protobuf.loadSync('system-test/fixtures/provinces.proto');
  const Province = root.lookupType('utilities.Province');

  // Create an event handler to handle messages
  let messageCount = 0;
  const messageHandler = async message => {
    // "Ack" (acknowledge receipt of) the message
    message.ack();

    // Get the schema metadata from the message.
    const schemaMetadata = Schema.metadataFromMessage(message.attributes);

    let result;
    switch (schemaMetadata.encoding) {
      case Encodings.Binary:
        result = Province.decode(message.data);
        break;
      case Encodings.Json:
        // This doesn't require decoding with the protobuf library,
        // since it's plain JSON. But you can still validate it against
        // your schema.
        result = JSON.parse(message.data.toString());
        console.log(`Validation of JSON: ${Province.verify(result)}`);
        break;
      default:
        console.log(`Unknown schema encoding: ${schemaMetadata.encoding}`);
        break;
    }

    console.log(`Received message ${message.id}:`);
    console.log(`\tData: ${JSON.stringify(result, null, 4)}`);
    console.log(`\tAttributes: ${JSON.stringify(message.attributes, null, 4)}`);
    messageCount += 1;
  };

  // Listen for new messages until timeout is hit
  subscription.on('message', messageHandler);

  setTimeout(() => {
    subscription.removeListener('message', messageHandler);
    console.log(`${messageCount} message(s) received.`);
  }, timeout * 1000);
}

PHP

Antes de tentar esse exemplo, siga as instruções de configuração do PHP em Guia de início rápido: como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Pub/Sub PHP.

Avro

Ver no GitHub Feedback

use Google\Cloud\PubSub\PubSubClient;

/**
 * Subscribe and pull messages using an AVRO schema.
 *
 * @param string $projectId
 * @param string $subscriptionId
 */
function subscribe_avro_records($projectId, $subscriptionId)
{
    $pubsub = new PubSubClient([
        'projectId' => $projectId,
    ]);

    $subscription = $pubsub->subscription($subscriptionId);
    $messages = $subscription->pull();

    foreach ($messages as $message) {
        $decodedMessageData = '';
        $encoding = $message->attribute('googclient_schemaencoding');
        switch ($encoding) {
            case 'BINARY':
                $ioReader = new \AvroStringIO(base64_decode($message->data()));
                $dataReader = new \AvroDataIOReader($ioReader, new \AvroIODatumReader());

                $decodedMessageData = json_encode($dataReader->data());
                break;
            case 'JSON':
                $decodedMessageData = $message->data();
                break;
        }

        printf('Received a %d-encoded message %s', $encoding, $decodedMessageData);
    }
}

Buffer de protocolo

Ver no GitHub Feedback

use Google\Cloud\PubSub\PubSubClient;

/**
 * Subscribe and pull messages using a protocol buffer schema.
 *
 * Relies on a proto message of the following form:
 * ```
 * syntax = "proto3";
 *
 * package utilities;
 *
 * message StateProto {
 *   string name = 1;
 *   string post_abbr = 2;
 * }
 * ```
 *
 * @param string $projectId
 * @param string $subscriptionId
 */
function subscribe_proto_messages($projectId, $subscriptionId)
{
    $pubsub = new PubSubClient([
        'projectId' => $projectId,
    ]);

    $subscription = $pubsub->subscription($subscriptionId);
    $messages = $subscription->pull();

    foreach ($messages as $message) {
        $decodedMessageData = '';
        $encoding = $message->attribute('googclient_schemaencoding');
        switch ($encoding) {
            case 'BINARY':
                $protobufMessage = new \Utilities\StateProto();
                $protobufMessage->mergeFromString($message->data());

                $decodedMessageData = $protobufMessage->serializeToJsonString();
                break;
            case 'JSON':
                $decodedMessageData = $message->data();
                break;
        }

        printf('Received a %d-encoded message %s', $encoding, $decodedMessageData);
    }
}

Python

Avro

Ver no GitHub Feedback

import avro
from avro.io import BinaryDecoder, DatumReader
from concurrent.futures import TimeoutError
import io
import json
from google.cloud.pubsub import SubscriberClient

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"
# avsc_file = "path/to/an/avro/schema/file/(.avsc)/formatted/in/json"
# Number of seconds the subscriber listens for messages
# timeout = 5.0

subscriber = SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

avro_schema = avro.schema.parse(open(avsc_file, "rb").read())

def callback(message):
    # Get the message serialization type.
    encoding = message.attributes.get("googclient_schemaencoding")
    # Deserialize the message data accordingly.
    if encoding == "BINARY":
        bout = io.BytesIO(message.data)
        decoder = BinaryDecoder(bout)
        reader = DatumReader(avro_schema)
        message_data = reader.read(decoder)
        print(f"Received a binary-encoded message:\n{message_data}")
    elif encoding == "JSON":
        message_data = json.loads(message.data)
        print(f"Received a JSON-encoded message:\n{message_data}")
    else:
        print(f"Received a message with no encoding:\n{message}")

    message.ack()

streaming_pull_future = subscriber.subscribe(subscription_path, callback=callback)
print(f"Listening for messages on {subscription_path}..\n")

# Wrap subscriber in a 'with' block to automatically call close() when done.
with subscriber:
    try:
        # When `timeout` is not set, result() will block indefinitely,
        # unless an exception occurs first.
        streaming_pull_future.result(timeout=timeout)
    except TimeoutError:
        streaming_pull_future.cancel()  # Trigger the shutdown.
        streaming_pull_future.result()  # Block until the shutdown is complete.

Buffer de protocolo

Ver no GitHub Feedback

from concurrent.futures import TimeoutError
from google.cloud.pubsub import SubscriberClient
from google.protobuf.json_format import Parse

from utilities import us_states_pb2

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"
# Number of seconds the subscriber listens for messages
# timeout = 5.0

subscriber = SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

# Instantiate a protoc-generated class defined in `us-states.proto`.
state = us_states_pb2.StateProto()

def callback(message):
    # Get the message serialization type.
    encoding = message.attributes.get("googclient_schemaencoding")
    # Deserialize the message data accordingly.
    if encoding == "BINARY":
        state.ParseFromString(message.data)
        print("Received a binary-encoded message:\n{state}")
    elif encoding == "JSON":
        Parse(message.data, state)
        print(f"Received a JSON-encoded message:\n{state}")
    else:
        print(f"Received a message with no encoding:\n{message}")

    message.ack()

streaming_pull_future = subscriber.subscribe(subscription_path, callback=callback)
print(f"Listening for messages on {subscription_path}..\n")

# Wrap subscriber in a 'with' block to automatically call close() when done.
with subscriber:
    try:
        # When `timeout` is not set, result() will block indefinitely,
        # unless an exception occurs first.
        streaming_pull_future.result(timeout=timeout)
    except TimeoutError:
        streaming_pull_future.cancel()  # Trigger the shutdown.
        streaming_pull_future.result()  # Block until the shutdown is complete.

Ruby

Avro

Ver no GitHub Feedback

# subscription_id = "your-subscription-id"
# avsc_file = "path/to/an/avro/schema/file/(.avsc)/formatted/in/json"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id

subscriber = subscription.listen do |received_message|
  encoding = received_message.attributes["googclient_schemaencoding"]
  case encoding
  when "BINARY"
    require "avro"
    avro_schema = Avro::Schema.parse File.read(avsc_file)
    buffer = StringIO.new received_message.data
    decoder = Avro::IO::BinaryDecoder.new buffer
    reader = Avro::IO::DatumReader.new avro_schema
    message_data = reader.read decoder
    puts "Received a binary-encoded message:\n#{message_data}"
  when "JSON"
    require "json"
    message_data = JSON.parse received_message.data
    puts "Received a JSON-encoded message:\n#{message_data}"
  else
    "Received a message with no encoding:\n#{received_message.message_id}"
  end
  received_message.acknowledge!
end

subscriber.start
# Let the main thread sleep for 60 seconds so the thread for listening
# messages does not quit
sleep 60
subscriber.stop.wait!

Buffer de protocolo

Ver no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"
require_relative "utilities/us-states_pb"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id

subscriber = subscription.listen do |received_message|
  encoding = received_message.attributes["googclient_schemaencoding"]
  case encoding
  when "BINARY"
    state = Utilities::StateProto.decode received_message.data
    puts "Received a binary-encoded message:\n#{state}"
  when "JSON"
    require "json"
    state = Utilities::StateProto.decode_json received_message.data
    puts "Received a JSON-encoded message:\n#{state}"
  else
    "Received a message with no encoding:\n#{received_message.message_id}"
  end
  received_message.acknowledge!
end

subscriber.start
# Let the main thread sleep for 60 seconds so the thread for listening
# messages does not quit
sleep 60
subscriber.stop.wait!

StreamingPull

O serviço Pub/Sub tem duas APIs para recuperar mensagens:

Sempre que possível, as bibliotecas de cliente do Cloud usam StreamingPull (em inglês) para máxima capacidade e menor latência. Embora seja possível que o uso da API StreamingPull diretamente nunca venha a acontecer, é importante entender algumas propriedades essenciais do StreamingPull e como ela difere do método Pull mais tradicional.

O método pull depende de um modelo de solicitação/resposta:

O cliente envia uma solicitação de mensagens ao servidor.
O servidor responde com zero ou mais mensagens e fecha a conexão.

A API do serviço StreamingPull depende de uma conexão bidirecional persistente para receber várias mensagens à medida que elas são disponibilizadas:

O cliente envia uma solicitação ao servidor para estabelecer uma conexão.
O servidor envia continuamente mensagens para o cliente conectado.
A conexão é eventualmente encerrada pelo cliente ou pelo servidor.

Você fornece um callback para o assinante, e ele executa o callback para cada mensagem de maneira assíncrona. Se um assinante receber mensagens com a mesma chave de ordenação, as bibliotecas de cliente executarão sequencialmente o retorno de chamada. O serviço do Pub/Sub entrega essas mensagens ao mesmo assinante, com base no melhor esforço.

StreamingPull tem uma taxa de erro de 100%, o que é esperado

Os streams do StreamingPull sempre fecham com um status não OK. Observe que, ao contrário das RPCs normais, o status aqui é simplesmente uma indicação de que o stream foi interrompido, não de que as solicitações estão falhando. Portanto, ainda que a API StreamingPull tenha uma taxa de erro de 100% que pareça surpreendente, ela foi projetada dessa maneira.

Como diagnosticar erros do StreamingPull

Como os streams do StreamingPull sempre terminam com um erro, não é útil examinar as métricas de encerramento de stream enquanto diagnostica erros. Em vez disso, concentre-se na métrica de operação de mensagem StreamingPull (subscription/streaming_pull_message_operation_count). Procure estes erros:

Erros FAILED_PRECONDITION podem ocorrer nos casos a seguir:
- Pub/Sub tenta descriptografar uma mensagem com uma chave desativada do Cloud KMS.
- As assinaturas podem ser temporariamente suspensas se houver mensagens no backlog da assinatura criptografadas com uma chave do Cloud KMS desativada.
Erros UNAVAILABLE

StreamingPull: como lidar com grandes backlogs de pequenas mensagens

A pilha gRPC StreamingPull está otimizada para alta capacidade e, portanto, armazena mensagens em buffers. Isso pode ter algumas consequências se você está tentando processar grandes backlogs de pequenas mensagens, em vez de um stream estável de novas mensagens. Nessas condições, você pode ver mensagens entregues várias vezes, e o balanço de carga delas pode não ser realizado com eficácia entre os clientes.

O buffer entre o serviço Pub/Sub e o espaço de usuário da biblioteca de cliente é de aproximadamente 10 MB. Para entender o impacto desse buffer no comportamento da biblioteca de cliente, analise este exemplo:

Há um backlog de 10.000 mensagens de 1 KB em uma assinatura.
Cada mensagem leva um segundo para ser processada sequencialmente por uma instância de cliente com um único thread.
A primeira instância de cliente para estabelecer uma conexão de StreamingPull com o serviço para essa inscrição preencherá seu buffer com todas as 10.000 mensagens.
Leva 10.000 segundos (quase três horas) para processar o buffer.
Nesse período, algumas das mensagens armazenadas em buffer excedem o prazo de confirmação e são reenviadas para o mesmo cliente, resultando em duplicatas.
Quando várias instâncias de cliente estão em execução, as mensagens presas no buffer de um cliente não ficam disponíveis para nenhuma instância de cliente.

Essa situação não ocorrerá se você usar o controle de fluxo para StreamingPull: o serviço nunca tem todos os 10 MB de mensagens por vez e, portanto, consegue balancear a carga de maneira eficaz entre vários assinantes. para começar.

Para resolver essa situação, use uma assinatura push ou a API Pull, atualmente disponível em algumas das bibliotecas de cliente do Cloud (consulte a seção Pull síncrono) e todas as bibliotecas de cliente da API. Para saber mais, consulte a documentação das bibliotecas de cliente.

Pull síncrono

Há casos em que o pull assíncrono não é adequado para seu aplicativo. Por exemplo, a lógica do aplicativo pode depender de um padrão de pesquisa para recuperar mensagens ou exigir um limite preciso em um número de mensagens recuperadas pelo cliente a qualquer momento. A compatibilidade do serviço com esses aplicativos depende do método de pull síncrono.

Veja a seguir um código de amostra para enviar por pull e confirmar um número fixo de mensagens:

C#

Veja no GitHub Feedback


using Google.Cloud.PubSub.V1;
using Grpc.Core;
using System;
using System.Linq;
using System.Text;
using System.Threading;

public class PullMessagesSyncSample
{
    public int PullMessagesSync(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);
        SubscriberServiceApiClient subscriberClient = SubscriberServiceApiClient.Create();
        int messageCount = 0;
        try
        {
            // Pull messages from server,
            // allowing an immediate response if there are no messages.
            PullResponse response = subscriberClient.Pull(subscriptionName, returnImmediately: false, maxMessages: 20);
            // Print out each received message.
            foreach (ReceivedMessage msg in response.ReceivedMessages)
            {
                string text = System.Text.Encoding.UTF8.GetString(msg.Message.Data.ToArray());
                Console.WriteLine($"Message {msg.Message.MessageId}: {text}");
                Interlocked.Increment(ref messageCount);
            }
            // If acknowledgement required, send to server.
            if (acknowledge && messageCount > 0)
            {
                subscriberClient.Acknowledge(subscriptionName, response.ReceivedMessages.Select(msg => msg.AckId));
            }
        }
        catch (RpcException ex) when (ex.Status.StatusCode == StatusCode.Unavailable)
        {
            // UNAVAILABLE due to too many concurrent pull requests pending for the given subscription.
        }
        return messageCount;
    }
}

Go

Veja no GitHub Feedback

import (
	"context"
	"fmt"
	"io"
	"time"

	"google.golang.org/grpc/codes"
	"google.golang.org/grpc/status"

	"cloud.google.com/go/pubsub"
)

func pullMsgsSync(w io.Writer, projectID, subID string) error {
	// projectID := "my-project-id"
	// subID := "my-sub"
	ctx := context.Background()
	client, err := pubsub.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("pubsub.NewClient: %v", err)
	}
	defer client.Close()

	sub := client.Subscription(subID)

	// Turn on synchronous mode. This makes the subscriber use the Pull RPC rather
	// than the StreamingPull RPC, which is useful for guaranteeing MaxOutstandingMessages,
	// the max number of messages the client will hold in memory at a time.
	sub.ReceiveSettings.Synchronous = true
	sub.ReceiveSettings.MaxOutstandingMessages = 10

	// Receive messages for 5 seconds.
	ctx, cancel := context.WithTimeout(ctx, 5*time.Second)
	defer cancel()

	// Create a channel to handle messages to as they come in.
	cm := make(chan *pubsub.Message)
	defer close(cm)
	// Handle individual messages in a goroutine.
	go func() {
		for msg := range cm {
			fmt.Fprintf(w, "Got message :%q\n", string(msg.Data))
			msg.Ack()
		}
	}()

	// Receive blocks until the passed in context is done.
	err = sub.Receive(ctx, func(ctx context.Context, msg *pubsub.Message) {
		cm <- msg
	})
	if err != nil && status.Code(err) != codes.Canceled {
		return fmt.Errorf("Receive: %v", err)
	}

	return nil
}

Java

Veja no GitHub Feedback


import com.google.cloud.pubsub.v1.stub.GrpcSubscriberStub;
import com.google.cloud.pubsub.v1.stub.SubscriberStub;
import com.google.cloud.pubsub.v1.stub.SubscriberStubSettings;
import com.google.pubsub.v1.AcknowledgeRequest;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PullRequest;
import com.google.pubsub.v1.PullResponse;
import com.google.pubsub.v1.ReceivedMessage;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class SubscribeSyncExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";
    Integer numOfMessages = 10;

    subscribeSyncExample(projectId, subscriptionId, numOfMessages);
  }

  public static void subscribeSyncExample(
      String projectId, String subscriptionId, Integer numOfMessages) throws IOException {
    SubscriberStubSettings subscriberStubSettings =
        SubscriberStubSettings.newBuilder()
            .setTransportChannelProvider(
                SubscriberStubSettings.defaultGrpcTransportProviderBuilder()
                    .setMaxInboundMessageSize(20 * 1024 * 1024) // 20MB (maximum message size).
                    .build())
            .build();

    try (SubscriberStub subscriber = GrpcSubscriberStub.create(subscriberStubSettings)) {
      String subscriptionName = ProjectSubscriptionName.format(projectId, subscriptionId);
      PullRequest pullRequest =
          PullRequest.newBuilder()
              .setMaxMessages(numOfMessages)
              .setSubscription(subscriptionName)
              .build();

      // Use pullCallable().futureCall to asynchronously perform this operation.
      PullResponse pullResponse = subscriber.pullCallable().call(pullRequest);
      List<String> ackIds = new ArrayList<>();
      for (ReceivedMessage message : pullResponse.getReceivedMessagesList()) {
        // Handle received message
        // ...
        ackIds.add(message.getAckId());
      }
      // Acknowledge received messages.
      AcknowledgeRequest acknowledgeRequest =
          AcknowledgeRequest.newBuilder()
              .setSubscription(subscriptionName)
              .addAllAckIds(ackIds)
              .build();

      // Use acknowledgeCallable().futureCall to asynchronously perform this operation.
      subscriber.acknowledgeCallable().call(acknowledgeRequest);
      System.out.println(pullResponse.getReceivedMessagesList());
    }
  }
}

Node.js

Veja no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const projectId = 'YOUR_PROJECT_ID';
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';

// Imports the Google Cloud client library. v1 is for the lower level
// proto access.
const {v1} = require('@google-cloud/pubsub');

// Creates a client; cache this for further use.
const subClient = new v1.SubscriberClient();

async function synchronousPull() {
  const formattedSubscription = subClient.subscriptionPath(
    projectId,
    subscriptionName
  );

  // The maximum number of messages returned for this request.
  // Pub/Sub may return fewer than the number specified.
  const request = {
    subscription: formattedSubscription,
    maxMessages: 10,
  };

  // The subscriber pulls a specified number of messages.
  const [response] = await subClient.pull(request);

  // Process the messages.
  const ackIds = [];
  for (const message of response.receivedMessages) {
    console.log(`Received message: ${message.message.data}`);
    ackIds.push(message.ackId);
  }

  if (ackIds.length !== 0) {
    // Acknowledge all of the messages. You could also acknowledge
    // these individually, but this is more efficient.
    const ackRequest = {
      subscription: formattedSubscription,
      ackIds: ackIds,
    };

    await subClient.acknowledge(ackRequest);
  }

  console.log('Done.');
}

synchronousPull().catch(console.error);

PHP

Veja no GitHub Feedback

use Google\Cloud\PubSub\PubSubClient;

/**
 * Pulls all Pub/Sub messages for a subscription.
 *
 * @param string $projectId  The Google project ID.
 * @param string $subscriptionName  The Pub/Sub subscription name.
 */
function pull_messages($projectId, $subscriptionName)
{
    $pubsub = new PubSubClient([
        'projectId' => $projectId,
    ]);
    $subscription = $pubsub->subscription($subscriptionName);
    foreach ($subscription->pull() as $message) {
        printf('Message: %s' . PHP_EOL, $message->data());
        // Acknowledge the Pub/Sub message has been received, so it will not be pulled multiple times.
        $subscription->acknowledge($message);
    }
}

Protocolo

Solicitação:

POST https://pubsub.googleapis.com/v1/projects/myproject/subscriptions/mysubscription:pull

{
  "returnImmediately": "false",
  "maxMessages": "1"
}

Resposta:

200 OK

{
  "receivedMessages": [{
    "ackId": "dQNNHlAbEGEIBERNK0EPKVgUWQYyODM2LwgRHFEZDDsLRk1SK...",
    "message": {
      "data": "SGVsbG8gQ2xvdWQgUHViL1N1YiEgSGVyZSBpcyBteSBtZXNzYWdlIQ==",
      "messageId": "19917247034"
    }
  }]
}

Solicitação:

POST https://pubsub.googleapis.com/v1/projects/myproject/subscriptions/mysubscription:acknowledge

{
  "ackIds": [
    "dQNNHlAbEGEIBERNK0EPKVgUWQYyODM2LwgRHFEZDDsLRk1SK..."
  ]
}

Python

Veja no GitHub Feedback

from google.api_core import retry
from google.cloud import pubsub_v1

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"

subscriber = pubsub_v1.SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

NUM_MESSAGES = 3

# Wrap the subscriber in a 'with' block to automatically call close() to
# close the underlying gRPC channel when done.
with subscriber:
    # The subscriber pulls a specific number of messages. The actual
    # number of messages pulled may be smaller than max_messages.
    response = subscriber.pull(
        request={"subscription": subscription_path, "max_messages": NUM_MESSAGES},
        retry=retry.Retry(deadline=300),
    )

    ack_ids = []
    for received_message in response.received_messages:
        print(f"Received: {received_message.message.data}.")
        ack_ids.append(received_message.ack_id)

    # Acknowledges the received messages so they will not be sent again.
    subscriber.acknowledge(
        request={"subscription": subscription_path, "ack_ids": ack_ids}
    )

    print(
        f"Received and acknowledged {len(response.received_messages)} messages from {subscription_path}."
    )

Ruby

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
subscription.pull(immediate: false).each do |message|
  puts "Message pulled: #{message.data}"
  message.acknowledge!
end

O Pub/Sub entrega uma lista de mensagens. Se a lista tiver várias mensagens, o Pub/Sub as ordena com a mesma chave de ordenação.

Para conseguir baixa latência de entrega de mensagens com pull síncrono, é importante ter muitas solicitações de pull pendentes ao mesmo tempo. À medida que a taxa de transferência do tópico aumenta, mais solicitações de pull são necessárias. Em geral, o pull assíncrono é indicado para aplicativos sensíveis à latência.

Pull síncrono com gerenciamento de lease

O processamento de uma mensagem individual pode exceder o prazo de confirmação pré-configurado, também conhecido como lease. Para evitar que essas mensagens sejam enviadas novamente, as bibliotecas de cliente permitem redefinir prazos de confirmação (exceto a biblioteca de cliente Go, que modifica automaticamente os prazos de confirmação de mensagens pesquisadas), conforme mostrado nas amostras abaixo:

C#

Veja no GitHub Feedback


using Google.Cloud.PubSub.V1;
using Grpc.Core;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;

public class PullMessageWithLeaseManagementSample
{
    public int PullMessageWithLeaseManagement(string projectId, string subscriptionId, bool acknowledge)
    {
        SubscriptionName subscriptionName = SubscriptionName.FromProjectSubscription(projectId, subscriptionId);
        SubscriberServiceApiClient subscriberClient = SubscriberServiceApiClient.Create();

        var ackIds = new List<string>();
        try
        {
            PullResponse response = subscriberClient.Pull(subscriptionName, returnImmediately: false, maxMessages: 20);

            // Print out each received message.
            foreach (ReceivedMessage msg in response.ReceivedMessages)
            {
                ackIds.Add(msg.AckId);
                string text = System.Text.Encoding.UTF8.GetString(msg.Message.Data.ToArray());
                Console.WriteLine($"Message {msg.Message.MessageId}: {text}");

                // Modify the ack deadline of each received message from the default 10 seconds to 30.
                // This prevents the server from redelivering the message after the default 10 seconds
                // have passed.
                subscriberClient.ModifyAckDeadline(subscriptionName, new List<string> { msg.AckId }, 30);
            }
            // If acknowledgement required, send to server.
            if (acknowledge && ackIds.Count > 0)
            {
                subscriberClient.Acknowledge(subscriptionName, ackIds);
            }
        }
        catch (RpcException ex) when (ex.Status.StatusCode == StatusCode.Unavailable)
        {
            // UNAVAILABLE due to too many concurrent pull requests pending for the given subscription.
        }
        return ackIds.Count;
    }
}

Java

Veja no GitHub Feedback


import com.google.cloud.pubsub.v1.stub.GrpcSubscriberStub;
import com.google.cloud.pubsub.v1.stub.SubscriberStub;
import com.google.cloud.pubsub.v1.stub.SubscriberStubSettings;
import com.google.pubsub.v1.AcknowledgeRequest;
import com.google.pubsub.v1.ModifyAckDeadlineRequest;
import com.google.pubsub.v1.ProjectSubscriptionName;
import com.google.pubsub.v1.PullRequest;
import com.google.pubsub.v1.PullResponse;
import com.google.pubsub.v1.ReceivedMessage;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class SubscribeSyncWithLeaseExample {
  public static void main(String... args) throws Exception {
    // TODO(developer): Replace these variables before running the sample.
    String projectId = "your-project-id";
    String subscriptionId = "your-subscription-id";
    Integer numOfMessages = 10;

    projectId = "tz-playground-bigdata";
    subscriptionId = "uno";

    subscribeSyncWithLeaseExample(projectId, subscriptionId, numOfMessages);
  }

  public static void subscribeSyncWithLeaseExample(
      String projectId, String subscriptionId, Integer numOfMessages)
      throws IOException, InterruptedException {
    SubscriberStubSettings subscriberStubSettings =
        SubscriberStubSettings.newBuilder()
            .setTransportChannelProvider(
                SubscriberStubSettings.defaultGrpcTransportProviderBuilder()
                    .setMaxInboundMessageSize(20 << 20) // 20 MB
                    .build())
            .build();

    try (SubscriberStub subscriber = GrpcSubscriberStub.create(subscriberStubSettings)) {

      String subscriptionName = ProjectSubscriptionName.format(projectId, subscriptionId);

      PullRequest pullRequest =
          PullRequest.newBuilder()
              .setMaxMessages(numOfMessages)
              .setSubscription(subscriptionName)
              .build();

      // Use pullCallable().futureCall to asynchronously perform this operation.
      PullResponse pullResponse = subscriber.pullCallable().call(pullRequest);

      List<String> ackIds = new ArrayList<>();

      for (ReceivedMessage message : pullResponse.getReceivedMessagesList()) {
        ackIds.add(message.getAckId());

        // Modify the ack deadline of each received message from the default 10 seconds to 30.
        // This prevents the server from redelivering the message after the default 10 seconds
        // have passed.
        ModifyAckDeadlineRequest modifyAckDeadlineRequest =
            ModifyAckDeadlineRequest.newBuilder()
                .setSubscription(subscriptionName)
                .addAckIds(message.getAckId())
                .setAckDeadlineSeconds(30)
                .build();

        subscriber.modifyAckDeadlineCallable().call(modifyAckDeadlineRequest);
      }

      // Acknowledge received messages.
      AcknowledgeRequest acknowledgeRequest =
          AcknowledgeRequest.newBuilder()
              .setSubscription(subscriptionName)
              .addAllAckIds(ackIds)
              .build();

      // Use acknowledgeCallable().futureCall to asynchronously perform this operation.
      subscriber.acknowledgeCallable().call(acknowledgeRequest);
      System.out.println(pullResponse.getReceivedMessagesList());
    }
  }
}

Node.js

Veja no GitHub Feedback

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const projectId = 'YOUR_PROJECT_ID';
// const subscriptionName = 'YOUR_SUBSCRIPTION_NAME';

// Imports the Google Cloud client library. v1 is for the lower level
// proto access.
const {v1} = require('@google-cloud/pubsub');

// Creates a client; cache this for further use.
const subClient = new v1.SubscriberClient();

async function synchronousPullWithLeaseManagement() {
  const formattedSubscription = subClient.subscriptionPath(
    projectId,
    subscriptionName
  );

  // The maximum number of messages returned for this request.
  // Pub/Sub may return fewer than the number specified.
  const maxMessages = 1;
  const newAckDeadlineSeconds = 30;
  const request = {
    subscription: formattedSubscription,
    maxMessages: maxMessages,
    allowExcessMessages: false,
  };

  let isProcessed = false;

  // The worker function is meant to be non-blocking. It starts a long-
  // running process, such as writing the message to a table, which may
  // take longer than the default 10-sec acknowledge deadline.
  function worker(message) {
    console.log(`Processing "${message.message.data}"...`);

    setTimeout(() => {
      console.log(`Finished procesing "${message.message.data}".`);
      isProcessed = true;
    }, 30000);
  }

  // The subscriber pulls a specified number of messages.
  const [response] = await subClient.pull(request);

  // Obtain the first message.
  const message = response.receivedMessages[0];

  // Send the message to the worker function.
  worker(message);

  let waiting = true;
  while (waiting) {
    await new Promise(r => setTimeout(r, 10000));
    // If the message has been processed..
    if (isProcessed) {
      const ackRequest = {
        subscription: formattedSubscription,
        ackIds: [message.ackId],
      };

      //..acknowledges the message.
      await subClient.acknowledge(ackRequest);
      console.log(`Acknowledged: "${message.message.data}".`);
      // Exit after the message is acknowledged.
      waiting = false;
      console.log('Done.');
    } else {
      // If the message is not yet processed..
      const modifyAckRequest = {
        subscription: formattedSubscription,
        ackIds: [message.ackId],
        ackDeadlineSeconds: newAckDeadlineSeconds,
      };

      //..reset its ack deadline.
      await subClient.modifyAckDeadline(modifyAckRequest);

      console.log(
        `Reset ack deadline for "${message.message.data}" for ${newAckDeadlineSeconds}s.`
      );
    }
  }
}

synchronousPullWithLeaseManagement().catch(console.error);

Python

Veja no GitHub Feedback

import logging
import multiprocessing
import sys
import time

from google.api_core import retry
from google.cloud import pubsub_v1

multiprocessing.log_to_stderr()
logger = multiprocessing.get_logger()
logger.setLevel(logging.INFO)
processes = dict()

# TODO(developer)
# project_id = "your-project-id"
# subscription_id = "your-subscription-id"

subscriber = pubsub_v1.SubscriberClient()
subscription_path = subscriber.subscription_path(project_id, subscription_id)

response = subscriber.pull(
    request={"subscription": subscription_path, "max_messages": 3},
    retry=retry.Retry(deadline=300),
)

# Start a process for each message based on its size modulo 10.
for message in response.received_messages:
    process = multiprocessing.Process(
        target=time.sleep, args=(sys.getsizeof(message) % 10,)
    )
    processes[process] = (message.ack_id, message.message.data)
    process.start()

while processes:
    # Take a break every second.
    if processes:
        time.sleep(1)

    for process in list(processes):
        ack_id, msg_data = processes[process]
        # If the process is running, reset the ack deadline.
        if process.is_alive():
            subscriber.modify_ack_deadline(
                request={
                    "subscription": subscription_path,
                    "ack_ids": [ack_id],
                    # Must be between 10 and 600.
                    "ack_deadline_seconds": 15,
                }
            )
            logger.info(f"Reset ack deadline for {msg_data}.")

        # If the process is complete, acknowledge the message.
        else:
            subscriber.acknowledge(
                request={"subscription": subscription_path, "ack_ids": [ack_id]}
            )
            logger.info(f"Acknowledged {msg_data}.")
            processes.pop(process)
print(
    f"Received and acknowledged {len(response.received_messages)} messages from {subscription_path}."
)

# Close the underlying gPRC channel. Alternatively, wrap subscriber in
# a 'with' block to automatically call close() when done.
subscriber.close()

Ruby

Veja no GitHub Feedback

# subscription_id = "your-subscription-id"
require "google/cloud/pubsub"

pubsub = Google::Cloud::Pubsub.new

subscription = pubsub.subscription subscription_id
new_ack_deadline = 30
processed = false

# The subscriber pulls a specified number of messages.
received_messages = subscription.pull immediate: false, max: 1

# Obtain the first message.
message = received_messages.first

# Send the message to a non-blocking worker that starts a long-running process, such as writing
# the message to a table, which may take longer than the default 10-sec acknowledge deadline.
Thread.new do
  sleep 15
  processed = true
  puts "Finished processing \"#{message.data}\"."
end

loop do
  sleep 1
  if processed
    # If the message has been processed, acknowledge the message.
    message.acknowledge!
    puts "Done."
    # Exit after the message is acknowledged.
    break
  else
    # If the message has not yet been processed, reset its ack deadline.
    message.modify_ack_deadline! new_ack_deadline
    puts "Reset ack deadline for \"#{message.data}\" for #{new_ack_deadline} seconds."
  end
end

Escalonamento

Pode ser necessário implementar um mecanismo de escalonamento no seu aplicativo de assinante para acompanhar o volume das mensagens. Como fazer isso depende do seu ambiente, mas ele geralmente será baseado nas métricas de backlog oferecidas por meio do Pacote de operações do Google Cloud do serviço de monitoramento. Para ver detalhes sobre como fazer isso para o Compute Engine, consulte Como escalonar com base em métricas do Cloud Monitoring.

Acesse a seção Pub/Sub da página lista de métricas do GCP para saber quais métricas podem ser monitoradas programaticamente.

Assim como ocorre com qualquer serviço distribuído, de vez em quando você terá de repetir cada solicitação.

Como lidar com duplicatas e forçar novas tentativas

Quando você não confirma uma mensagem antes do vencimento do prazo de confirmação, o Pub/Sub reenvia a mensagem. Como resultado, o Pub/Sub pode enviar mensagens duplicadas. Use o pacote de operações do Google Cloud para monitorar as operações de confirmação com o código de resposta expired para detectar essa condição. Para conseguir esses dados, selecione a métrica Confirmar operações de mensagens e agrupe ou filtre pelo rótulo response_code. Observe que response_code é um rótulo do sistema em uma métrica. Ele não é uma métrica.

Use o Stackdriver para pesquisar prazos de confirmação de mensagens vencidos

Para reduzir a taxa de duplicação, prolongue o prazo da mensagem.

As bibliotecas de cliente prolongam o prazo automaticamente, mas há limites padrão de quanto o prazo pode ser prolongado.
Use o método modifyAckDeadline para estender o prazo de confirmação se você estiver criando sua própria biblioteca de cliente.

Como alternativa, para forçar o Pub/Suba repetir uma mensagem, defina modifyAckDeadline para 0.