Google 会使用 AI 技术将内容翻译成您偏好的语言。AI 翻译可能包含错误。

重试请求

发布失败通常是由客户端瓶颈导致的，例如服务 CPU 不足，线程运行状况不佳或网络拥塞。发布者重试政策定义了 Pub/Sub 尝试传送消息的次数以及每次尝试之间的时间长度。

本文档介绍了如何将重试请求与发布到主题的消息搭配使用。

准备工作

在配置发布工作流之前，请确保您已完成以下任务：

了解主题和发布工作流。
创建一个主题。

所需角色

如需获得重试向主题发送消息请求所需的权限，请让您的管理员向您授予主题的Pub/Sub 发布者 (roles/pubsub.publisher) IAM 角色。如需详细了解如何授予角色，请参阅管理对项目、文件夹和组织的访问权限。

您也可以通过自定义角色或其他预定义角色来获取所需的权限。

您需要额外的权限才能创建或更新主题和订阅。

重试请求简介

重试设置控制 Pub/Sub 客户端库重试发布请求的方式。客户端库具有以下重试设置：

初始请求超时：客户端库停止等待初始发布请求完成前的时间。
重试延迟：请求超时后，客户端库等待请求重试的时间。
总超时：客户端库停止重试发布请求后的时间量。

如需重试发布请求，初始请求超时必须短于总超时。例如，如果您使用的是指数退避算法，客户端库会按如下所示计算请求超时和重试延迟：

每个发布请求后，请求超时会以请求超时倍数增加，最长为请求超时上限。
每次重试后，重试延迟时间会以重试延迟倍数增加，最长为重试延迟上限。

重试消息请求

在发布过程中，您可能会看到暂时性或永久性发布失败。对于暂时性错误，您通常无需采取任何特殊操作，因为 Pub/Sub 会自动重试消息。

当发布操作成功，但发布者客户端未及时收到发布响应时，也可能会发生错误。在这种情况下，系统也会重试发布操作。因此，您可能会收到两条内容相同但消息 ID 不同的消息。

如果出现持续性错误，请考虑在发布过程之外实现适当的操作，以避免 Pub/Sub 负载过重。

如果发布失败，系统会自动重试，但无法保证能够重试的错误除外。此示例代码演示了如何使用自定义重试设置创建发布者（请注意，并非所有客户端库都支持自定义重试设置；请参阅适用于您所选语言的API 参考文档）：

C++

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 C++ 设置说明进行操作。如需了解详情，请参阅 Pub/Sub C++ API 参考文档。

namespace pubsub = ::google::cloud::pubsub;
using ::google::cloud::future;
using ::google::cloud::Options;
using ::google::cloud::StatusOr;
[](std::string project_id, std::string topic_id) {
  auto topic = pubsub::Topic(std::move(project_id), std::move(topic_id));
  // By default a publisher will retry for 60 seconds, with an initial backoff
  // of 100ms, a maximum backoff of 60 seconds, and the backoff will grow by
  // 30% after each attempt. This changes those defaults.
  auto publisher = pubsub::Publisher(pubsub::MakePublisherConnection(
      std::move(topic),
      Options{}
          .set<pubsub::RetryPolicyOption>(
              pubsub::LimitedTimeRetryPolicy(
                  /*maximum_duration=*/std::chrono::minutes(10))
                  .clone())
          .set<pubsub::BackoffPolicyOption>(
              pubsub::ExponentialBackoffPolicy(
                  /*initial_delay=*/std::chrono::milliseconds(200),
                  /*maximum_delay=*/std::chrono::seconds(45),
                  /*scaling=*/2.0)
                  .clone())));

  std::vector<future<bool>> done;
  for (char const* data : {"1", "2", "3", "go!"}) {
    done.push_back(
        publisher.Publish(pubsub::MessageBuilder().SetData(data).Build())
            .then([](future<StatusOr<std::string>> f) {
              return f.get().ok();
            }));
  }
  publisher.Flush();
  int count = 0;
  for (auto& f : done) {
    if (f.get()) ++count;
  }
  std::cout << count << " messages sent successfully\n";
}

C#

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 C# 设置说明进行操作。如需了解详情，请参阅 Pub/Sub C# API 参考文档。


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

public class PublishMessageWithRetrySettingsAsyncSample
{
    public async Task PublishMessageWithRetrySettingsAsync(string projectId, string topicId, string messageText)
    {
        TopicName topicName = TopicName.FromProjectTopic(projectId, topicId);
        // Retry settings control how the publisher handles retry-able failures
        var maxAttempts = 3;
        var initialBackoff = TimeSpan.FromMilliseconds(110); // default: 100 ms
        var maxBackoff = TimeSpan.FromSeconds(70); // default : 60 seconds
        var backoffMultiplier = 1.3; // default: 1.0
        var totalTimeout = TimeSpan.FromSeconds(100); // default: 600 seconds

        var publisher = await new PublisherClientBuilder
        {
            TopicName = topicName,
            ApiSettings = new PublisherServiceApiSettings
            {
                PublishSettings = CallSettings.FromRetry(RetrySettings.FromExponentialBackoff(
                               maxAttempts: maxAttempts,
                               initialBackoff: initialBackoff,
                               maxBackoff: maxBackoff,
                               backoffMultiplier: backoffMultiplier,
                               retryFilter: RetrySettings.FilterForStatusCodes(StatusCode.Unavailable)))
                       .WithTimeout(totalTimeout)
            }
        }.BuildAsync();
        string message = await publisher.PublishAsync(messageText);
        Console.WriteLine($"Published message {message}");
        // PublisherClient instance should be shutdown after use.
        // The TimeSpan specifies for how long to attempt to publish locally queued messages.
        await publisher.ShutdownAsync(TimeSpan.FromSeconds(15));
    }
}

Go

以下示例使用 Go Pub/Sub 客户端库的主要版本 (v2)。如果您仍在使用 v1 库，请参阅 v2 迁移指南。如需查看 v1 代码示例列表，请参阅已废弃的代码示例。

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 Go 设置说明进行操作。如需了解详情，请参阅 Pub/Sub Go API 参考文档。

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

	"cloud.google.com/go/pubsub/v2"
	vkit "cloud.google.com/go/pubsub/v2/apiv1"
	gax "github.com/googleapis/gax-go/v2"
	"google.golang.org/grpc/codes"
)

func publishWithRetrySettings(w io.Writer, projectID, topicID, msg string) error {
	// projectID := "my-project-id"
	// topicID := "my-topic"
	// msg := "Hello World"
	ctx := context.Background()

	config := &pubsub.ClientConfig{
		TopicAdminCallOptions: &vkit.TopicAdminCallOptions{
			Publish: []gax.CallOption{
				gax.WithRetry(func() gax.Retryer {
					return gax.OnCodes([]codes.Code{
						codes.Aborted,
						codes.Canceled,
						codes.Internal,
						codes.ResourceExhausted,
						codes.Unknown,
						codes.Unavailable,
						codes.DeadlineExceeded,
					}, gax.Backoff{
						Initial:    250 * time.Millisecond, // default 100 milliseconds
						Max:        60 * time.Second,       // default 60 seconds
						Multiplier: 1.45,                   // default 1.3
					})
				}),
			},
		},
	}

	client, err := pubsub.NewClientWithConfig(ctx, projectID, config)
	if err != nil {
		return fmt.Errorf("pubsub: NewClient: %w", err)
	}
	defer client.Close()

	// client.Publisher can be passed a topic ID (e.g. "my-topic") or
	// a fully qualified name (e.g. "projects/my-project/topics/my-topic").
	// If a topic ID is provided, the project ID from the client is used.
	// Reuse this publisher for all publish calls to send messages in batches.
	publisher := client.Publisher(topicID)
	result := publisher.Publish(ctx, &pubsub.Message{
		Data: []byte(msg),
	})
	// Block until the result is returned and a server-generated
	// ID is returned for the published message.
	id, err := result.Get(ctx)
	if err != nil {
		return fmt.Errorf("pubsub: result.Get: %w", err)
	}
	fmt.Fprintf(w, "Published a message; msg ID: %v\n", id)
	return nil
}

Java

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 Java 设置说明进行操作。如需了解详情，请参阅 Pub/Sub Java API 参考文档。


import com.google.api.core.ApiFuture;
import com.google.api.gax.retrying.RetrySettings;
import com.google.cloud.pubsub.v1.Publisher;
import com.google.protobuf.ByteString;
import com.google.pubsub.v1.PubsubMessage;
import com.google.pubsub.v1.TopicName;
import java.io.IOException;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.TimeUnit;
import org.threeten.bp.Duration;

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

    publishWithRetrySettingsExample(projectId, topicId);
  }

  public static void publishWithRetrySettingsExample(String projectId, String topicId)
      throws IOException, ExecutionException, InterruptedException {
    TopicName topicName = TopicName.of(projectId, topicId);
    Publisher publisher = null;

    try {
      // Retry settings control how the publisher handles retry-able failures
      Duration initialRetryDelay = Duration.ofMillis(100); // default: 100 ms
      double retryDelayMultiplier = 2.0; // back off for repeated failures, default: 1.3
      Duration maxRetryDelay = Duration.ofSeconds(60); // default : 60 seconds
      Duration initialRpcTimeout = Duration.ofSeconds(1); // default: 5 seconds
      double rpcTimeoutMultiplier = 1.0; // default: 1.0
      Duration maxRpcTimeout = Duration.ofSeconds(600); // default: 600 seconds
      Duration totalTimeout = Duration.ofSeconds(600); // default: 600 seconds

      RetrySettings retrySettings =
          RetrySettings.newBuilder()
              .setInitialRetryDelay(initialRetryDelay)
              .setRetryDelayMultiplier(retryDelayMultiplier)
              .setMaxRetryDelay(maxRetryDelay)
              .setInitialRpcTimeout(initialRpcTimeout)
              .setRpcTimeoutMultiplier(rpcTimeoutMultiplier)
              .setMaxRpcTimeout(maxRpcTimeout)
              .setTotalTimeout(totalTimeout)
              .build();

      // Create a publisher instance with default settings bound to the topic
      publisher = Publisher.newBuilder(topicName).setRetrySettings(retrySettings).build();

      String message = "first message";
      ByteString data = ByteString.copyFromUtf8(message);
      PubsubMessage pubsubMessage = PubsubMessage.newBuilder().setData(data).build();

      // Once published, returns a server-assigned message id (unique within the topic)
      ApiFuture<String> messageIdFuture = publisher.publish(pubsubMessage);
      String messageId = messageIdFuture.get();
      System.out.println("Published a message with retry settings: " + messageId);

    } finally {
      if (publisher != null) {
        // When finished with the publisher, shutdown to free up resources.
        publisher.shutdown();
        publisher.awaitTermination(1, TimeUnit.MINUTES);
      }
    }
  }
}

Node.js

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 Node.js 设置说明进行操作。如需了解详情，请参阅Pub/Sub Node.js API 参考文档。

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const projectId = 'YOUR_PROJECT_ID'
// const topicNameOrId = 'YOUR_TOPIC_NAME_OR_ID';
// const data = JSON.stringify({foo: 'bar'});

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

async function publishWithRetrySettings(topicNameOrId, data) {
  const pubsubClient = new PubSub();

  // Retry settings control how the publisher handles retryable failures. Default values are shown.
  // The `retryCodes` array determines which grpc errors will trigger an automatic retry.
  // The `backoffSettings` object lets you specify the behaviour of retries over time.
  //
  // Reference this document to see the current defaults for publishing:
  // https://github.com/googleapis/nodejs-pubsub/blob/6e2c28a9298a49dc1b194ce747ff5258c8df6deb/src/v1/publisher_client_config.json#L59
  //
  // Please note that _all_ items must be included when passing these settings to topic().
  // Otherwise, unpredictable (incorrect) defaults may be assumed.
  const retrySettings = {
    retryCodes: [
      10, // 'ABORTED'
      1, // 'CANCELLED',
      4, // 'DEADLINE_EXCEEDED'
      13, // 'INTERNAL'
      8, // 'RESOURCE_EXHAUSTED'
      14, // 'UNAVAILABLE'
      2, // 'UNKNOWN'
    ],
    backoffSettings: {
      // The initial delay time, in milliseconds, between the completion
      // of the first failed request and the initiation of the first retrying request.
      initialRetryDelayMillis: 100,
      // The multiplier by which to increase the delay time between the completion
      // of failed requests, and the initiation of the subsequent retrying request.
      retryDelayMultiplier: 4,
      // The maximum delay time, in milliseconds, between requests.
      // When this value is reached, retryDelayMultiplier will no longer be used to increase delay time.
      maxRetryDelayMillis: 60000,
      // The initial timeout parameter to the request.
      initialRpcTimeoutMillis: 60000,
      // The multiplier by which to increase the timeout parameter between failed requests.
      rpcTimeoutMultiplier: 1.0,
      // The maximum timeout parameter, in milliseconds, for a request. When this value is reached,
      // rpcTimeoutMultiplier will no longer be used to increase the timeout.
      maxRpcTimeoutMillis: 60000,
      // The total time, in milliseconds, starting from when the initial request is sent,
      // after which an error will be returned, regardless of the retrying attempts made meanwhile.
      totalTimeoutMillis: 600000,
    },
  };

  // Cache topic objects (publishers) and reuse them.
  const topic = pubsubClient.topic(topicNameOrId, {
    gaxOpts: {
      retry: retrySettings,
    },
  });

  // Publishes the message as a string, e.g. "Hello, world!" or JSON.stringify(someObject)
  const dataBuffer = Buffer.from(data);
  const messageId = await topic.publishMessage({data: dataBuffer});
  console.log(`Message ${messageId} published.`);
}

Node.js

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 Node.js 设置说明进行操作。如需了解详情，请参阅Pub/Sub Node.js API 参考文档。

/**
 * TODO(developer): Uncomment these variables before running the sample.
 */
// const projectId = 'YOUR_PROJECT_ID'
// const topicNameOrId = 'YOUR_TOPIC_NAME_OR_ID';
// const data = JSON.stringify({foo: 'bar'});

// Imports the Google Cloud client library. v1 is for the lower level
// proto access.
import {PubSub} from '@google-cloud/pubsub';

async function publishWithRetrySettings(topicNameOrId: string, data: string) {
  const pubsubClient = new PubSub();

  // Retry settings control how the publisher handles retryable failures. Default values are shown.
  // The `retryCodes` array determines which grpc errors will trigger an automatic retry.
  // The `backoffSettings` object lets you specify the behaviour of retries over time.
  //
  // Reference this document to see the current defaults for publishing:
  // https://github.com/googleapis/nodejs-pubsub/blob/6e2c28a9298a49dc1b194ce747ff5258c8df6deb/src/v1/publisher_client_config.json#L59
  //
  // Please note that _all_ items must be included when passing these settings to topic().
  // Otherwise, unpredictable (incorrect) defaults may be assumed.
  const retrySettings = {
    retryCodes: [
      10, // 'ABORTED'
      1, // 'CANCELLED',
      4, // 'DEADLINE_EXCEEDED'
      13, // 'INTERNAL'
      8, // 'RESOURCE_EXHAUSTED'
      14, // 'UNAVAILABLE'
      2, // 'UNKNOWN'
    ],
    backoffSettings: {
      // The initial delay time, in milliseconds, between the completion
      // of the first failed request and the initiation of the first retrying request.
      initialRetryDelayMillis: 100,
      // The multiplier by which to increase the delay time between the completion
      // of failed requests, and the initiation of the subsequent retrying request.
      retryDelayMultiplier: 4,
      // The maximum delay time, in milliseconds, between requests.
      // When this value is reached, retryDelayMultiplier will no longer be used to increase delay time.
      maxRetryDelayMillis: 60000,
      // The initial timeout parameter to the request.
      initialRpcTimeoutMillis: 60000,
      // The multiplier by which to increase the timeout parameter between failed requests.
      rpcTimeoutMultiplier: 1.0,
      // The maximum timeout parameter, in milliseconds, for a request. When this value is reached,
      // rpcTimeoutMultiplier will no longer be used to increase the timeout.
      maxRpcTimeoutMillis: 60000,
      // The total time, in milliseconds, starting from when the initial request is sent,
      // after which an error will be returned, regardless of the retrying attempts made meanwhile.
      totalTimeoutMillis: 600000,
    },
  };

  // Cache topic objects (publishers) and reuse them.
  const topic = pubsubClient.topic(topicNameOrId, {
    gaxOpts: {
      retry: retrySettings,
    },
  });

  // Publishes the message as a string, e.g. "Hello, world!" or JSON.stringify(someObject)
  const dataBuffer = Buffer.from(data);
  const messageId = await topic.publishMessage({data: dataBuffer});
  console.log(`Message ${messageId} published.`);
}

Python

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 Python 设置说明进行操作。如需了解详情，请参阅 Pub/Sub Python API 参考文档。

from google import api_core
from google.cloud import pubsub_v1

# TODO(developer)
# project_id = "your-project-id"
# topic_id = "your-topic-id"

# Configure the retry settings. Defaults shown in comments are values applied
# by the library by default, instead of default values in the Retry object.
custom_retry = api_core.retry.Retry(
    initial=0.250,  # seconds (default: 0.1)
    maximum=90.0,  # seconds (default: 60.0)
    multiplier=1.45,  # default: 1.3
    deadline=300.0,  # seconds (default: 60.0)
    predicate=api_core.retry.if_exception_type(
        api_core.exceptions.Aborted,
        api_core.exceptions.DeadlineExceeded,
        api_core.exceptions.InternalServerError,
        api_core.exceptions.ResourceExhausted,
        api_core.exceptions.ServiceUnavailable,
        api_core.exceptions.Unknown,
        api_core.exceptions.Cancelled,
    ),
)

publisher = pubsub_v1.PublisherClient()
topic_path = publisher.topic_path(project_id, topic_id)

for n in range(1, 10):
    data_str = f"Message number {n}"
    # Data must be a bytestring
    data = data_str.encode("utf-8")
    future = publisher.publish(topic=topic_path, data=data, retry=custom_retry)
    print(future.result())

print(f"Published messages with retry settings to {topic_path}.")

使用排序键重试请求

假设您有一个发布者客户端。您正在使用 Pub/Sub 客户端库为同一排序键 A 发布消息 1、2 和 3。现在，假设发布者客户端在 RPC 截止时间到期之前未收到消息 1 的发布响应。消息 1 必须重新发布。如果您假设消息 2 仅在消息 1 成功完成后发布，那么订阅者客户端收到的消息序列将变为 1、1、2 和 3。每条发布的消息都有自己的消息 ID。从订阅者客户端的角度来看，发布了四条消息，前两条消息的内容相同。

使用排序键重试发布请求也可能会因批处理设置而变得复杂。客户端库会将消息批量处理在一起，以便更高效地发布。继续使用前面的示例，并假设消息 1 和 2 批量处理在一起。此批处理将作为单个请求发送到服务器。如果服务器未能及时返回响应，发布者客户端会重试此批处理中的两条消息。因此，订阅者客户端可能会收到消息 1、2、1、2 和 3。如果您使用 Pub/Sub 客户端库按顺序发布消息，并且发布操作失败，则该服务会使同一排序键上的所有剩余消息的发布操作失败。然后，发布者客户端可以决定执行以下任一操作：

按顺序重新发布所有失败的消息
按顺序重新发布失败消息的子集
发布一组新消息

如果出现不可重试的错误，则客户端库无法发布消息，并会停止发布带有相同排序键的其他消息。例如，当发布者将消息发送到不存在的主题时，即会发生不可重试的错误。要继续发布带用相同排序键的消息，请调用一个方法以继续发布，然后开始重新发布。

以下示例显示了如何继续发布带有相同排序键的消息。

C++

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 C++ 设置说明进行操作。如需了解详情，请参阅 Pub/Sub C++ API 参考文档。

namespace pubsub = ::google::cloud::pubsub;
using ::google::cloud::future;
using ::google::cloud::StatusOr;
[](pubsub::Publisher publisher) {
  struct SampleData {
    std::string ordering_key;
    std::string data;
  } data[] = {
      {"key1", "message1"}, {"key2", "message2"}, {"key1", "message3"},
      {"key1", "message4"}, {"key1", "message5"},
  };
  std::vector<future<void>> done;
  for (auto& datum : data) {
    auto const& da = datum;  // workaround MSVC lambda capture confusion
    auto handler = [da, publisher](future<StatusOr<std::string>> f) mutable {
      auto const msg = da.ordering_key + "#" + da.data;
      auto id = f.get();
      if (!id) {
        std::cout << "An error has occurred publishing " << msg << "\n";
        publisher.ResumePublish(da.ordering_key);
        return;
      }
      std::cout << "Message " << msg << " published as id=" << *id << "\n";
    };
    done.push_back(
        publisher
            .Publish(pubsub::MessageBuilder{}
                         .SetData("Hello World! [" + datum.data + "]")
                         .SetOrderingKey(datum.ordering_key)
                         .Build())
            .then(handler));
  }
  publisher.Flush();
  // Block until all the messages are published (optional)
  for (auto& f : done) f.get();
}

C#

在尝试此示例之前，请按照《快速入门：使用客户端库》中的 C# 设置说明进行操作。如需了解详情，请参阅 Pub/Sub C# API 参考文档。


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

public class ResumePublishSample
{
    public async Task<int> PublishOrderedMessagesAsync(string projectId, string topicId, IEnumerable<(string, string)> keysAndMessages)
    {
        TopicName topicName = TopicName.FromProjectTopic(projectId, topicId);

        var customSettings = new PublisherClient.Settings
        {
            EnableMessageOrdering = true
        };

        PublisherClient publisher = await new PublisherClientBuilder
        {
            TopicName = topicName,
            Settings = customSettings
        }.BuildAsync();

        int publishedMessageCount = 0;
        var publishTasks = keysAndMessages.Select(async keyAndMessage =>
        {
            try
            {
                string message = await publisher.PublishAsync(keyAndMessage.Item1, keyAndMessage.Item2);
                Console.WriteLine($"Published message {message}");

重试请求 使用集合让一切井井有条 根据您的偏好保存内容并对其进行分类。

准备工作

所需角色

重试请求简介

重试消息请求

C++

C#

Go

Java

Node.js

Node.js

Python

使用排序键重试请求

C++

C#

重试请求